Now in its 14th season, ROBOTICON Tampa Bay 2026, is the Foundation for Community Driven Innovation’s signature Manufacturing Month event. The event...
Location: AMRoC Fab Lab
Tampa AI Week blends conference-level insights with the energy of a high-impact meetup. As a leading Florida AI Conference, Tampa AI Week brings together operators, executives, developers, and innovators for practical sessions, community events, and meaningful connections across the Tampa AI ecosystem.
Learn how to use Instagram & Facebook, leverage AI tools, and create content that boosts your SEO and attracts more customers.
Location: Entrepreneur Collaborative Center
Join us for our October 2026 Meetup! Here are some details for this event: Hosted By: Dr. Tom Stablein, Program Director Global Online M.S. AIBA Program Muma College of Business, University of South Florida Purpose: For most businesses and organizations, data analytics and business intelligence is used to support mandatory business functions. For data scientists, the tools and technologies are ever-changing. Join local experts in discussing current trends and topics in the analytics field, inclu
Location: 4202 E Fowler Ave
Join us for our October 2026 Meetup! Here are some details for this event: **Hosted By:** Dr. Tom Stablein, Program Director [Global Online M.S. AIBA Program](https://www.usf.edu/business/graduate/ms-ai-business-analytics-global/index.aspx) Muma College of Business, University of South Florida **Purpose:** For most businesses and organizations, data analytics and business intelligence is used to support mandatory business functions. For data scientists, the tools and technologies are ever-changing. Join local experts in discussing current trends and topics in the analytics field, including AI and ML, data science, business analytics, systems design, database management, data mining, and more.. **Agenda:** 6:00pm-6:30pm – refreshments and networking 6:30pm-7:30pm – presentation and discussion 7:30pm-8:00pm – networking [building closes at 8pm] **Location:** The [Muma College of Business](https://www.usf.edu/business/) at the University of South Florida is located at **4202 E. Fowler Ave., Tampa, FL 33620** **Parking:** Please park in Reserved Lot 22D which is across from Muma College of Business. You can enter the lot near the Yuengling Center. **RSVP:** Please RSVP early, since seating is limited to 60 attendees. If you intend to bring a guest, please have them RSVP separately, so we can plan refreshments. **Cost:** These events are provided at no cost to the local data analytics community, courtesy of the Global Online Master's of Science in Artificial Intelligence (AI) and Business Analytics program at the Muma College of Business, University of South Florida. For more info, visit [usf.to/global-msaiba](https://usf.to/global-msaiba).
Location: 4202 E Fowler Ave
Tampa AI Week blends conference-level insights with the energy of a high-impact meetup. As a leading Florida AI Conference, Tampa AI Week brings together operators, executives, developers, and innovators for practical sessions, community events, and meaningful connections across the Tampa AI ecosystem.
Location: Museum of Science & Innovation - MOSI
Orlando's 19th Annual Day of Data (SQL Saturday) is a FREE one-day training event for IT professionals and students interested in data platforms.
A Look at Airtable and Supabase for Websites - ZOOM based event Speaker: Rodney Biddle This is an event to specifically talk Database sources for Vibe (Or Traditional) coded websites. Airtable? Supabase? Google Sheets or FireBase? We will touch on all four and also discuss JSON files as data sources. The primary focus will be Airtable and Supabase though. I ues both heavily but the conditions differ extremely. This will be a discussion as I demonstrate my use and we discuss why and restrictions.
This month we will be doing a use demonstration of Hermes and taking a look at Open Claw as well. Through the use of Hermes over the last few months I have been able to automate many functions recently including being a part of the efforts. We will also be reviewing a new website that I have been developing focused on the Job Applicant market in Tampa Bay and Central Florida. Also - Due to low attendance recently, we will be tracking member attendance, specificaly tracking no-shows. If you need
Location: Hillsborough County ECC
Join us in Tampa to explore how AI is shaking up CRM and boosting customer connections live!
Location: Stewart Title Company - Tampa
A free conference for students, professionals, academia and tech enthusiasts that brought 336 attendees last year.
Location: Marshall Student Center, University of South Florida, Tampa, FL
"The sociotechnical imaginaries of AI produce technological promises subject to a cycle of hype, disillusionment, and criticism. These future prognostications evolve as AI’s capabilities expand, even as they result in a process of routinization, yielding artistic approaches rooted in the technology’s failure to realize these espoused promises. The sociotechnical imaginary is a useful lens to examine how technological promises have produced AI-enhanced music practices from a pragmatism that continually reconciles with a techno-utopian futurity. The persuasiveness of AI’s evolving technological promise has shaped discourses, expectations, and disillusion towards the present and future of musical creation. The myth of the intelligent machine—an assemblage or apparatus that cannot just automate human labor but possesses creative agency—contains philosophical and creative implications for musical intelligence in an age when an AI ecology consolidates the roles involved in music composition and performance while accelerating the process of its production. This talk examines three case studies that traverse the development of the AI’s sociotechnical imaginaries, relating them to music creation. These instruments and platforms—Minsky and Friedkin’s Triadex Muse, Puckett and Zicarelli’s Max/Msp~/Jitter, and the current commercial end-to-end AI-driven music platforms for hyper-reproduction, were developed in response to a contextualized cycle of hype, disillusion, and criticism. The routinization brought upon by the disillusion voiced towards the preceding technological promises greatly informed the creative scope and ambition in the following epoch of sociotechnological imaginaries. These case studies will be used to critique AI-driven music production under the latest era of the sociotechnical imaginary, demonstrating that contemporary technological promises will also be subject to an inevitable process of routinization. Yet, there are opportunities for bridging the previous imaginaries with the present promises, which can foster a routinization process that centers the performer as the real-time interlocutor between an evolving technological assemblage and their own musical intelligence, agency, and creativity. Rather than focusing on AI’s rote technical functionality, these case studies and supporting theoretical sources will be used to both critique and confront these existing technological promises, and leverage them into contemplating the future of human-centered technological mediation in music creation. "
Speaker: Steven Lewis
Location: ENG 4
Live! 360 Tech Con brings together IT pros, developers, data and security experts, enterprise and AI engineers, and technology leaders to sharpen skills, share best practices, and stay ahead of what’s next. Join us in Orlando this November as we put the Future in Focus—deepening your AI expertise, fueling innovation, and propelling your career forward.
Join us for our November 2026 Meetup! Here are some details for this event: **Hosted By:** Dr. Tom Stablein, Program Director [Global Online M.S. AIBA Program](https://www.usf.edu/business/graduate/ms-ai-business-analytics-global/index.aspx) Muma College of Business, University of South Florida **Purpose:** For most businesses and organizations, data analytics and business intelligence is used to support mandatory business functions. For data scientists, the tools and technologies are ever-changing. Join local experts in discussing current trends and topics in the analytics field, including AI and ML, data science, business analytics, systems design, database management, data mining, and more.. **Agenda:** 6:00pm-6:30pm – refreshments and networking 6:30pm-7:30pm – presentation and discussion 7:30pm-8:00pm – networking [building closes at 8pm] **Location:** Entrepreneur Collaborative Center 2101 East Palm Avenue, Tampa FL 33605 **Parking:** Please do not park in front of the ECC facility. \*\*Free parking is available in the lot located directly North of the ECC building.\*\* **RSVP:** Please RSVP early, since seating is limited to 60 attendees. If you intend to bring a guest, please have them RSVP separately, so we can plan refreshments. **Cost:** These events are provided at no cost to the local data analytics community, courtesy of the Global Online Master's of Science in Artificial Intelligence (AI) and Business Analytics program at the Muma College of Business, University of South Florida. For more info, visit [usf.to/global-msaiba](https://usf.to/global-msaiba).
Location: Entrepreneur Collaborative Center
Join **Disrupt the Bay and Masterminds Tampa Bay for TECHAPALOOZA** [Registration](https://www.eventbrite.com/e/techapalooza-tickets-2002221970813) **\*\*RSVP - Be sure to use the Eventbrite Registration link at the top of the description to secure your spot\*\*** An evening bringing together leaders, innovators, and Tampa Bay’s growing AI and technology community. Our panel will cut through the AI hype and explore **how AI should actually be implemented**—from strategy and real-world applications to building solutions that create measurable impact. Connect directly with executives, technologists, founders, and AI professionals through **speed networking and open networking**, while enjoying food and beverages. Whether you’re already building with AI or trying to understand where it fits within your organization, TECHAPALOOZA is about **real conversations, practical insight, and meaningful connections with the people shaping Tampa Bay’s AI future.** ### TECHAPALOOZA \| EVENT AGENDA **November 19\, 2026 \| 5:00–8:00 PM \| SOFWERX** **5:00–5:30 PM — Doors Open + Food & Beverages** Check in, grab a drink, meet Tampa Bay’s AI and technology community, and get connected. **5:30–6:00 PM — Welcome to TECHAPALOOZA** Opening remarks from **Disrupt the Bay × Masterminds Tampa Bay** and introduction of our featured speakers. **6:00–6:30 PM — Featured AI Panel: From AI Hype to Real-World Implementation** A candid conversation about where AI is actually delivering value, where organizations are getting it wrong, and what leaders need to consider around adoption, strategy, governance, security and execution. Featuring: **Nick Palomba — Microsoft** **John O'reilly — Google** **Jennifer Bryk — RebelEdge Advisory** **René Clayton — Byte Strategy AI** **Brian Andrews - Sofwerx** **6:30–6:45 PM — Audience Q&A** **6:45–7:15 PM — TECHAPALOOZA Speed Networking ⚡** Fast-paced introductions designed to connect founders, executives, technologists, AI practitioners and innovators across Tampa Bay. **7:15–8:00 PM — Open Networking + Food & Beverages** Keep the conversations going, make meaningful connections and meet the people helping shape Tampa Bay’s technology ecosystem. **\*\*RSVP - Be sure to use the Eventbrite Registration link at the top of the description to secure your spot\*\***
Location: SOFWERX
Join us for our December 2026 Meetup! Here are some details for this event: **Hosted By:** Dr. Tom Stablein, Program Director [Global Online M.S. AIBA Program](https://www.usf.edu/business/graduate/ms-ai-business-analytics-global/index.aspx) Muma College of Business, University of South Florida **Purpose:** For most businesses and organizations, data analytics and business intelligence is used to support mandatory business functions. For data scientists, the tools and technologies are ever-changing. Join local experts in discussing current trends and topics in the analytics field, including AI and ML, data science, business analytics, systems design, database management, data mining, and more.. **Agenda:** 6:00pm-6:30pm – refreshments and networking 6:30pm-7:30pm – presentation and discussion 7:30pm-8:00pm – networking [building closes at 8pm] **Location:** Entrepreneur Collaborative Center 2101 East Palm Avenue, Tampa FL 33605 **Parking:** Please do not park in front of the ECC facility. \*\*Free parking is available in the lot located directly North of the ECC building.\*\* **RSVP:** Please RSVP early, since seating is limited to 60 attendees. If you intend to bring a guest, please have them RSVP separately, so we can plan refreshments. **Cost:** These events are provided at no cost to the local data analytics community, courtesy of the Global Online Master's of Science in Artificial Intelligence (AI) and Business Analytics program at the Muma College of Business, University of South Florida. For more info, visit [usf.to/global-msaiba](https://usf.to/global-msaiba).
Location: Entrepreneur Collaborative Center
Large language models are shifting from text predictors to agents that operate on the web and within software systems. This transition amplifies safety risks across dialogue, the environments agents control, and compliance with real-world policies. This talk presents ThinkGuard, a reasoning-based guardrail trained via mission-focused distillation to acquire deliberative thinking without costly manual annotation. By reasoning over goals, constraints, and latent hazards, ThinkGuard generalizes to implicit, complex, and previously unseen risks. I will also outline guardrail extensions that enforce policy compliance for web and system agents, and omnimodal guardrails that vet and steer interactions involving images, audio, and video. Together, these techniques transform guardrails from brittle filters into adaptive, explanatory safety layers that preserve utility while measurably reducing failure in agentic workflows.
Speaker: Muhao Chen
Despite rapid advances in artificial intelligence, many clinical AI models fail to achieve real-world healthcare impact. The primary barrier is often not model performance, but misalignment between AI methods, clinically meaningful questions, and access to appropriate real-world data. This talk presents a practical roadmap for computer science and engineering investigators interested in conducting high-impact clinical AI research at scale. We introduce a structured evidence-gap discovery framework that combines systematic reviews and clinical practice guidelines with AI-assisted extraction and human validation to identify clinically important, measurable, and fundable research questions. These gaps are then mapped to appropriate AI tasks�including prediction, natural language processing, computer vision, and multimodal modeling�and to real-world clinical, administrative, registry, and social determinant datasets. The session also reviews realistic data access pathways, governance considerations, and collaboration models. Attendees will leave with a concrete framework for moving from idea to dataset to fundable clinical AI study, grounded in real-world clinical needs and data constraints.
Speaker: Ricardo Pietrobon
This presentation will explore recent advances in controllable image generation using diffusion models, with a focus on achieving precise alignment between generated images and structured visual conditions such as segmentation masks, edges, depth maps, and human poses. The talk will begin by outlining the limitations of existing conditional diffusion methods, which often lack explicit objectives to enforce fine-grained controllability. I will then introduce two recent approaches that address this challenge through consistency- and preference-based learning. The first leverages pixel-level cycle consistency and efficient reward fine-tuning to strengthen adherence to input controls without incurring the high computational cost of full diffusion backpropagation. The second reformulates controllable generation as a condition-level preference optimization problem, enabling robust optimization across all diffusion timesteps. Together, these methods demonstrate significant improvements in controllability while preserving image quality, highlighting new directions for reliable and scalable control in generative vision models.
Speaker: Chen Chen
Large-scale geospatial understanding from satellite imagery remains less mature than natural-image vision. Current practice often relies on narrow, task-specific pipelines, and there is limited work on truly task-agnostic foundation models that remain robust under domain shift, geographic variation, and diverse real-world scenes. We advance remote-sensing foundation-model development and evaluation along three axes. First, we build self-supervised multispectral models for satellite image classification by adapting common SSL methods (BYOL, Barlow Twins, DINO, MoCo, SimSiam) to multispectral inputs. We pretrain on EuroSAT and BigEarthNet and evaluate cross-dataset transfer. The resulting multispectral checkpoints, especially those trained on BigEarthNet achieve strong transfer despite using only a fraction of the data typically used in ImageNet-scale training. Second, we benchmark eight ViT-based geospatial foundation models (including SatMAE/Scale-MAE/Cross-Scale MAE, the Geospatial Foundation Model, and supervised multi-task pretraining) via end-to-end fine-tuning on four downstream tasks: classification, segmentation, oriented object detection, and change detection. Performance is highly task-dependent: models are competitive for classification and segmentation but underperform for change detection and oriented detection, and cross-dataset generalization degrades when imagery comes from different satellites. Across experiments, pretraining data quality (spatial resolution and scene diversity) matters more than dataset size, and scaling from ViT-Base to ViT-Large yields ~1% gains. Finally, we introduce Hierarchical Attention Distillation (HAD) to transfer multi-scale spatial and semantic representations from a large teacher to a compact student using unlabeled data, reducing parameters to less than 5% of state-of-the-art solutions while retaining competitive performance.
Speaker: Gilbert Rotich
Current augmented reality (AR) headset applications often rely on predefined gesture sets that fail to align with users' natural expectations, leading to unintuitive experiences. To address this gap, we conducted two separate open-ended elicitation studies with 30 adults and 20 children (ages 9-12). Participants were tasked with replicating 17 different actions (e.g., moving, rotating) on a 3D virtual cube using any interaction modality they preferred.�Our findings reveal critical differences and similarities across both participant groups. Adults predominantly default to direct hand manipulation (comprising 82%�of first choices) , but demonstrate a strong preference for interaction flexibility, with 52%�of secondary proposals switching to a different modality. Their mental models are largely grounded in real-world experiences, naive physics, and pop culture. Conversely, children expect to engage their entire bodies (e.g., kicking, running) and frequently incorporate external physical or imagined objects as tools. Furthermore, children distinctively construct narratives around their interactions to contextualize abstract tasks. By comparing these distinct mental models, this talk highlights the necessity of moving beyond static, unimodal gesture sets. We will discuss actionable design recommendations for future AR headset applications, emphasizing the need for multimodal flexibility, whole-body tracking, and physics-aware virtual environments to accommodate the diverse, intuitive expectations of all users.
Speaker: Alejandro
Miami JVM Group (MJUG) *Join us for an exclusive in-person session where we'll explore **Kiro:** an **AI-powered developer platform** designed to help you turn ideas and specs into real, working code in minutes.* ***Kiro doesn't just autocomplete your code,*** *it **understands your project specs, c
Location: Miami JVM Group (MJUG)
Conversational artificial intelligence (C-AI) has the potential to scale speech-language therapy, yet current systems often fail when exposed to the variability and error patterns characteristic of disordered speech. This reflects a key limitation in the clinical development pipeline, where AI systems are evaluated under idealized conditions without preclinical testing in realistic impairment-rich communicative contexts. This talk introduces Agent-Based Conversational Dialogue (ABCD), a simulation-based framework that models impaired speech dialogue as a preclinical testbed for conversational AI in communication disorders. ABCD integrates agent-based modeling with large language models to simulate dynamic, multi-turn interactions between a virtual patient agent with communication disorders (e.g., aphasia) and a virtual clinician. This enables rapid, large-scale testing of AI systems without requiring human participants. Using this framework, two studies are presented. Study 1 demonstrates feasibility, showing that AI clinicians maintain high coherence and therapeutic appropriateness despite errorful input. Study 2 demonstrates that response accessibility is a tunable parameter influenced by prompting strategies, model architecture, and reasoning mode. Beyond evaluation, ABCD enables generation of synthetic disordered speech data by parameterizing linguistic error types, severity levels, and multimodal perturbations. These datasets may support robust, inclusive AI training. Together, ABCD establishes a scalable, simulation-first paradigm for developing safe and clinically effective conversational AI.
Speaker: Gerald C. Imaezue
Tampa AI Applications Meetup Group ONLINE EVENT - ZOOM CALL We are going to take a look at Manus.ai as a prompt-based tool that can be used for automating tasks. I currently use it for gathering news stories on two websites and am using it to collect event data for a third. It is prompt based and ca
Location: Tampa AI Applications Meetup Group
GDG Tampa Bay Join us for an exciting hands-on workshop where we'll dive into the world of Flutter development! In just 75 minutes, you'll learn how to create a single-screen Flutter app featuring a mesmerizing 3D-like orb. Watch as the app transforms to reflect the vibes entered by users, such as "
Location: GDG Tampa Bay
Miami JVM Group (MJUG) This session explores building AI-related solutions with Java, leveraging popular open-source frameworks such as Hibernate Vector, Spring AI, and LangChain4J. We’ll explore design patterns such as RAG, agentic AI, and vector search for storing, querying, and retrieving vector
Location: Miami JVM Group (MJUG)
Data Analytics & AI - Tampa Bay Join us for our April 2026 Meetup! Here are some details for this event: **Hosted By:** Dr. Tom Stablein, Program Director [Global Online M.S. AIBA Program]( Muma College of Business, University of South Florida **Purpose:** For most businesses and organizations, data
Location: Data Analytics & AI - Tampa Bay
Jacksonville JAVA User Group (JaxJUG) This session explores building AI-related solutions with Java, leveraging popular open-source frameworks such as Hibernate Vector, Spring AI, and LangChain4J. We’ll explore design patterns such as RAG, agentic AI, and vector search to store, query, and retrieve
Location: Jacksonville JAVA User Group (JaxJUG)
Orlando JUG (Java User Group) \#\# GenAI apps with LangChain4j and MCP * This session, designed for software engineers, will explore how to create GenAI apps with LangChain4j and MCP (Model Context Protocol), facilitating standardized interactions with external tools. Through practical examples, we’
Location: Orlando JUG (Java User Group)
Virtual Java User Group **Please register [HERE]( **vJUG Connect** is teaming up with the **London Java Community** for an evening all about AI and modern Java. Join Rod Johnson, creator of the Spring Framework, as he explores building JVM agents with Embabel, bridging Java and Generative AI for the
Location: Virtual Java User Group
Large language models (LLMs) are increasingly being integrated into hardware design verification workflows, particularly for SystemVerilog Assertion (SVA) generation. However, existing approaches primarily emphasize syntactic correctness and formal verification for evaluation, while the semantic alignment between generated assertions and natural language design specifications remains difficult to quantify and ensure. This gap introduces risks of hallucinated or vacuous assertions, reducing confidence and increasing debugging effort, especially in the absence of golden RTL. Bridging this gap requires a method that enforces semantic alignment throughout the generation process rather than a post-generation step.�This talk will begin with a review of the current state of the art SVA generation methodologies and their limitations. To address these challenges, we present SpecAlign, a framework for evaluating and refining LLM-generated SVAs with a novel approach that explicitly enforces semantic alignment between assertions and the specification. SpecAlign iteratively refines and provides evidence for natural language properties and SVAs with the goal of improving consistency, explainability, and confidence. This is achieved with two evidence- and reasoning-based loops that classify properties and SVAs based on semantic alignment and generate structured feedback and evidence. Additionally, SpecAlign provides a quantitative alignment score to measure semantic alignment across iterations, providing an alternative to traditional evaluation methods that rely on formal verification and golden RTL. The talk concludes with a discussion of future research directions in LLM-based SVA generation and the broader role of LLMs in hardware verification.
Speaker: Jaime Imperial
Tampa AI Applications Meetup Group This is ONLINE In preparation for the Vibe Coding event coming up on the 15th, I'm going to hold an Online event where I cover the preparation of the Airtable Database we will be using for our vibe coding project on the 15th. Airtable is a good light-duty database
Location: Tampa AI Applications Meetup Group
Tampa Bay Agile Agile is everywhere — but is it actually working? Industry data suggests adoption has spread wide while depth has plateaued, and many practitioners are feeling the tension between rising expectations and real-world constraints. In this session, we'll dig into some of the trends shapi
Location: Tampa Bay Agile
Tampa AI Applications Meetup Group A Very powerful use of AI tools includes "Vibe Coding". This is prompt coding to design a small to medium sized application. Vibe coding is not perfect - but the simpler the application or website, the easier it is requiring fewer skills. Complex applications it ma
Location: Tampa AI Applications Meetup Group
Tampa Open Claw Set Up Function This group comes together to share their experiences with Ai bots. We have a wide array of experience and those that wish to learn and share their actions will be in good company. Entrepreneurs have had wonderful success implementing Ai agents. We hope that people com
Location: Tampa Open Claw Set Up Function
Sensing is a crucial component of autonomous systems, providing the necessary information to interact with and understand the environment. However, environmental conditions often limit the effectiveness of sensing capabilities, making perception inherently uncertain. This talk will address the challenges of underwater sensing and perception, emphasizing the need for autonomous systems to manage uncertainties arising from dynamic and unknown environments. Drawing from lessons learned in a previous project on underwater object recognition, I will discuss the role of planning algorithms in active sensing to reduce perception uncertainties. I will also introduce ongoing research efforts that integrate planning and perception, focusing on how these methods can enhance the performance of autonomous systems in challenging underwater settings.
Speaker: Jane Sine
Tampa Devs Join us as we celebrate the sucess of speaker Ryan Clements, a long time Tampa Devs member, speaker and founder of Byte Bot. Ryan will share a story on how he and his team **deleted $600K in Software Costs (and the hard lessons learned along the way)**. Which involved replacing an expensi
Location: Tampa Devs
Tech Success Network Learn how to craft a killer technical resume that will land you interviews and impress hiring managers! Technology is a vast industry with so many diverse specializations. Now more than ever, it's critical that your resume represents not just your professional experience, but yo
Location: Tech Success Network
Tampa Machine Learning Join me in discussing the paper — scaling laws for neural language models.
Location: Tampa Machine Learning
Large Language Models (LLMs) are a key part of modern artificial intelligence, used in tasks ranging from complex reasoning to everyday interactions. As demand for better performance grows, they are made smarter through scaling, using more data, larger models, and more recently, Test-Time Scaling (TTS). TTS uses extra compute at inference time to generate better responses. While this improves output quality and reasoning, its safety risks are still not well understood. In this talk, I will explain how LLMs work, how scaling improves them, and why these improvements can also introduce new risks. I will begin with an overview of LLMs and scaling methods, focusing on TTS approaches such as Best-of-N. These methods can be understood through three main components: how responses are generated, how they are scored, and how much computation is used. Next, I will connect traditional adversarial attacks on LLMs to new risks introduced by TTS. I will then present our work, �Less Diverse, Less Safe�, which identifies a new failure mode related to diversity in generated responses. I will show that this safety concern arises across different TTS methods and models and demonstrate that the threat can be transferred to closed-source systems and can bypass strong safety guardrails. I conclude by highlighting the need for safety-aware TTS design as an important step toward building safer and more reliable AI systems.
Speaker: Shahriar Kabir Nahin
Tampa AI Applications Meetup Group Our Monthly in-Person meeting. Topic TBD - Feel free to submit suggestions in our discussion area. We will post the topic 30 days out.
Location: Tampa AI Applications Meetup Group
The conference provides an international platform for professionals, researchers, academicians, and industry experts involved in Artificial Intelligence and Emerging Technologies to exchange knowledge and gain insights into the latest developments,
Explore generative AI and AI agent skills.
Where business professionals can level up their skills and connect with peers in a dynamic, hands-on learning environment.
Join us for specialists aiconnect networking in Tampa on 10 Jul 2026, 5 PM EDT. Connect with professionals!
Build AI fluency and apply AI responsibly in real-world and workplace contexts.
A casual gathering of high-performing, non-conforming women founders, tech workers, VCs, and hype girls This month’s Tampa Chapter meeting is special as we’re partnering with Tampa Bay Tech Week to celebrate 8-1-3 Day during - a community-wide celebration of the people, companies, and ideas driving our region’s thriving innovation ecosystem. It’s the perfect opportunity to expand your network, exchange ideas, and build meaningful relationships with women across industries who believe that collaboration is more powerful than competition.
Regardless of what you are Vibe coding with - This is Important. Claude Code, Bolt, Lovable, Cursor, Replet, Base64... They all have the same weaknesses. No documentation. Weak database development. Exposing credentials. Not enforcing database RLS. No documentation. Poor database design leading to possible collapse under heavy use. SEO Invisible. Poor or no Social Media capabilities.
Get invited to meet Tampa technology employers hiring now.
Can Code Communities monthly webinars help make AI accesible to everyone covering topics like LLM chatbots, prompt engineering, agentic systems and more.
For most businesses and organizations, data analytics and business intelligence is used to support mandatory business functions. For data scientists, the tools and technologies are ever-changing. Join local experts in discussing current trends and topics in the analytics field, including AI and ML, data science, business analytics, systems design, database management, data mining, and more..
Join local experts in discussing current trends and topics in the analytics field, including AI and ML, data science, business analytics, systems design, database management, data mining, and more
Speaker: Ashish Bhutada
Location: 2101 E Palm Ave · Tampa, FL
Microsoft AZURE AI Foundry - A Project Manager's Perspective Presentation Overview Travis has attended multiple AI-focused sessions at the ECC and has found significant value in learning from the developers, engineers, and AI practitioners within the Tampa AI Applications Meetup community. After several recent encounters with Microsoft Azure AI Foundry, Travis began exploring the platform and recognized an opportunity to introduce an enterprise-level Microsoft AI development and delivery environ
Reinforcement learning (RL) promises agents that can improve through experience using only the supervision provided by a reward function. In robotics, however, real-world interaction is expensive, dense rewards are difficult to specify, and sparse rewards often lead to sample-inefficient learning. In this talk, I explore whether imperfect physics priors can nevertheless provide useful structure for RL. Focusing on goal-conditioned RL (GCRL) for robot navigation and manipulation, I show how control-theoretic principles can serve as inductive biases for policy learning. We first connect goal-reaching value functions to time-to-go fields and interpret the Eikonal equation as a simplified surrogate for the Hamilton-Jacobi-Bellman equation, motivating Eikonal-informed value learning for challenging navigation problems. We then study its interaction with the quasimetric structure typical of GCRL. Their combination reveals the limitations of imposing these priors globally and motivates a hierarchical formulation. This becomes particularly important in contact-rich manipulation, where hybrid dynamics induce state-dependent controllability and motivate applying physics-informed structure only to appropriate state representations.
Speaker: Vittorio Giammarino
Most AI coding assistants are sealed cloud platforms heavy on dashboards and light on local control. Enter pi.dev—an open-source, terminal-based agent harness that strips away the bloat to give local LLMs direct access to a minimal toolset (read, write, edit, bash). In this talk, we’ll explore how to harness this minimalist architecture to build a local "mini-software factory." I’ll show you how to feed a structured application specification to pi.dev and watch it autonomously architect, generat
Automate your social media content from research to posting - whether you're just getting started or ready to build a fully custom AI content machine.
How Leading Employers Are Solving the Talent Shortage Through Apprenticeships The workforce is changing faster than ever. Between AI, evolving skill requirements, an aging workforce, and increasing competition for experienced talent, employers are ...
As large language models scale to trillions of parameters, hardware constraints on memory and compute become severe physical bottlenecks. A key driver of this inefficiency is that neural network layers are not created equal: some layers do the heavy lifting to reduce errors, while others are nearly redundant. Existing methods attempt to manage this, but they rely on gradient-based trial-and-error heuristics that ignore the local "curvature" of the learning landscape. These heuristics offer no guarantees and are computationally expensive, leading to a suboptimal distribution of resources and performance. To solve this, we take inspiration from information theory, especially the Minimum Description Length principle, to design a framework that acts as a mathematically rigorous budget manager for LLMs and any other AI models. By calculating a curvature-adjusted score for each layer, our system assesses a layer's true reducible risk and its actual potential to improve the model. It then uses a single convex optimization program to automatically drive two capacity allocation decisions: allocating extra computing capacity to the high-performing layers or pruning away wasted parameters from the low-performing ones. Tested on models like Mistral-7B and Gemma-7B, with benchmarks which include CoLA, MRPC, CommonsenseQA, etc. This budget-feasible approach replaces empirical guesswork with provable optimization guarantees, maximizing efficiency strictly within a global hardware limit.
Speaker: Theophilus Amaefuna
Location: ENG 4
Tampa QA Meetup × Tampa JUG Present × Tampa Bay Tech & Beer AI-assisted coding is quickly becoming the primary way applications are built. While it allows developers to create complex applications and features with only a few prompts, ambiguous language and AI hallucinations can introduce bugs, hidden assumptions, and unintended behavior. Simply asking an AI coding agent to “write tests” isn’t enough to guarantee correctness. Billy Korando is a Developer Advocate at Oracle and a member of the Ja
Location: Hillsborough Community College - Dale Mabry Campus, Tampa, FL
**Tampa QA Meetup × Tampa JUG Present × Tampa Bay Tech & Beer** AI-assisted coding is quickly becoming the primary way applications are built. While it allows developers to create complex applications and features with only a few prompts, ambiguous language and AI hallucinations can introduce bugs, hidden assumptions, and unintended behavior. Simply asking an AI coding agent to “write tests” isn’t enough to guarantee correctness. **[Billy Korando](https://billykorando.com/)** is a Developer Advocate at [Oracle](https://www.oracle.com/) and a member of the Java Platform Group’s Developer Relations team. **What to Expect** Billy will walk through three distinct phases of testing AI-generated code—from defining expected business behavior and covering edge cases to uncovering bugs and risks you may not have anticipated. 🎯 **Highlights** * **Known knowns:** Use specification-driven development to clearly define the expected, business-relevant behavior of your application. * **Known unknowns:** Apply property-based testing to explore edge cases and avoid overfitting and unintended biases. * **Unknown unknowns:** Use static analysis and mutation testing to uncover issues that developers and AI agents may overlook. * **More reliable AI-generated code:** Move beyond simply asking an AI coding agent to “write tests.” * **Stronger test coverage:** Combine complementary testing methods to identify both expected and unexpected failures. * **A reusable testing framework:** Learn an approach that can evolve alongside rapidly changing AI coding tools. **Agenda** **5:45 PM–6:15 PM:** Food and networking **6:15 PM–7:15 PM:** Presentation **7:15 PM–7:45 PM:** Closing and raffle **7:45 PM onward:** Continue the conversation at a nearby bar or restaurant **Location** The event will take place in **DTEC 427.** Please see the attached image for the building and parking locations. **Sponsor** Thank you to our sponsor, **iDelsoft**, for its generous support. **iDelsoft** helps high-growth startups build and scale engineering teams through a network of more than 30,000 vetted developers across 150+ technologies. Hiring, or exploring your next opportunity? Visit [idelsoft.com](https://idelsoft.com/).
Location: Hillsborough College - Dale Mabry Campus
Large Language Models (LLMs) now exhibit remarkable growth in reasoning capabilities through test-time scaling (TTS). TTS uses additional compute during inference by generating thinking tokens. While TTS has significantly enhanced the reasoning capabilities of LLMs, these improvements come at the cost of increased inference time and computational demands. Model pruning provides a promising way to reduce these computational requirements by removing redundant parameters from the model. However, previous work has shown that naively pruning reasoning models can lead to severe degradation in performance. In this talk, I will show that, unlike structured pruning approaches that past work has explored in the context of reasoning and TTS, unstructured pruning methods can retain reasoning performance comparable to the full model, and at times even exceeding it. I will conclude by discussing what these findings reveal about reasoning models and their broader implications for building more efficient AI systems without sacrificing capability.
Speaker: Ocean Monjour
Location: ENG 4
Large Language Models (LLMs) now exhibit remarkable growth in reasoning capabilities through test-time scaling (TTS). TTS uses additional compute during inference by generating thinking tokens. While TTS has significantly enhanced the reasoning capabilities of LLMs, these improvements come at the cost of increased inference time and computational demands. Model pruning provides a promising way to reduce these computational requirements by removing redundant parameters from the model. However, previous work has shown that naively pruning reasoning models can lead to severe degradation in performance. In this talk, I will show that, unlike structured pruning approaches that past work has explored in the context of reasoning and TTS, unstructured pruning methods can retain reasoning performance comparable to the full model, and at times even exceeding it. I will conclude by discussing what these findings reveal about reasoning models and their broader implications for building more efficient AI systems without sacrificing capability.
Speaker: Ocean Monjur
Location: ENG 4
Learn how to use AI practically in your business. Understanding the fundamentals to choosing and implementing the right AI tools and work
Location: Lesley "Les" Miller Jr. All People's Community Park & Life Center
Cloud migration is no longer the finish line. Today's organizations are under pressure to reduce technical debt, modernize legacy applications, lower licensing costs, and prepare their environments for AI-driven development. Everyone is talking about AI-generated code, but what about the millions of lines of legacy code already running your business? In this session, we'll explore how AWS Transform can help organizations modernize applications, reduce technical debt, and turn legacy systems into
Location: Entrepreneur Collaborative Center (ECC), Tampa, FL
How Leading Employers Are Solving the Talent Shortage Through Apprenticeships The workforce is changing faster than ever. Between AI, evolving skill requirements, an aging workforce, and increasing competition for experienced talent, employers are ...
Location: Tampa, FL