AI Tinkerers - Seattle: Practical Agents Hackathon

Nikita Shamgunov is the founder and CEO of Neon, a serverless Postgres platform launched in 2021. He's developing technology that offers instant provisioning, autoscaling, and data branching to simplify database management for developers. Previously, Nikita co-founded MemSQL, scaling it to $40M ARR. He's an experienced tech entrepreneur and early-stage investor, focusing on deep technology and enterprise software.

Developing Neon, a serverless Postgres platform featuring instant provisioning, autoscaling, and data branching. Technical work includes building pg_embedding, an open-source extension for HNSW-based vector search, and optimizing AI-enabled indexing and distributed data systems.

Co-Founder & CEO of CopilotKit🪁 - open-source infra for building fully customized AI Copilots into any product (thousands of weekly installs, 8k+ GitHub stars). Leading & building dev infra my entire career (CopilotKit, Meta, Doximity). Past life: 2 physics degrees @UPenn.

n/a

Atai Barkai is currently leading the development of CopilotKit, an open-source infrastructure designed for integrating fully customized AI copilots into various products. The project has gained significant traction, with thousands of weekly installs and over 8,000 stars on GitHub. His work involves building developer infrastructure and enhancing AI solutions, leveraging his expertise in software development and media SDKs. Specific technologies and frameworks used in these projects have not been

Kyle Corbitt is the co-founder and CEO of OpenPipe, the RL post-training company. OpenPipe has trained thousands of customer models for both enterprises and tech-forward startups. Before founding OpenPipe, Kyle led the Startup School team at Y Combinator, which was responsible for the product and content that YC produces for early-stage companies. Prior to that he worked as an engineer at Google and studied ML at school.

RL and agents

https://github.com/OpenPipe/ART

I am a founding engineer at OpenPipe, an AI-based Seattle startup. My interests lie in AI-based startups tackling challenging problems in the realms of LLMs, fine-tuning, RAG, and QA. With a background encompassing research contributions to innovative systems like Prompt2Model and publications in prestigious conferences (EMNLP, ACL), I have also co-founded projects such as an AI-guided journaling assistant and a debate club. My industrial experience includes leading projects at Oracle and contributing to patent efforts at Wysa. Ultimately, my goal is to establish an AI-based startup focused on positive societal impact, leveraging my diverse experiences in research, entrepreneurship, and industry.

I am interested in: - People fine-tuning LLMs - People building consumer facing LLM apps - ML evaluations

https://openpipe.ai/

I completed my Master’s in Machine Learning at Texas A&M University and then spent about four years at Amazon as an Applied Scientist. During that time, I worked with the Multimodal team at Alexa and later with the Fashion Science team in Amazon Retail, developing computer vision models that are now part of various Amazon products.

Machine learning

Developing models for better application-logs understanding and code understanding.

Technical advisor to startups (preseed and seed) at AI2 Incubator.

Current projects using AI, LLMs: what are being built, how engineers build, best practices.

Agents

Founder, AI Tinkerers. Prior: Crowd Cow (>$50M ARR), MediaPiston (acq. Upwork), Snapvine (acq. Whitepages). Former co-founder/President, Madrona Venture Labs. UW CSE, MIT Sloan. Building the global AI Tinkerers community and platform.

Early-stage AI builders, agentic dev tools, applied LLM products, and technical founding teams. Open to intros for seed-stage teams and exceptional builders.

Building AI Tinkerers: events + platform (talks, hackathons, matchmaking, job board). Also advising a small set of seed-stage teams on hiring, distribution, and fundraising.

Founding Engineer at OpenPipe, where I build fine-tuning tools that help companies deploy faster, smaller, and more efficient LLMs. I focus on reinforcement learning, autonomous agents, model distillation, and evaluation. I also work on automating fine-tuning workflows and making models better aligned with real-world needs.

LLM, fine-tuning, agents, automation.

OpenPipe.ai

João Moura is the CEO of crewAI, with 20 years of experience. He previously studied at NYU Stern School of Business, FIAP, and General Assembly.

Building crewAI, a framework for orchestrating role-based AI agents. Developing multi-agent systems using Python to automate complex workflows. Focused on agentic process design, live-code templates, and scaling collaborative AI ecosystems for engineering and data science teams.

I’m a Technical Support Engineer who’s passionate about harnessing artificial intelligence to elevate the way we care for our customers.

Artificial intelligence, machine learning models, natural language understanding, real-time analytics

Building and integrating an LLM to handle support tickets.