Hackathon Team Chat - AI CTO Assistant

Role: Co-Creator, UI Finalization & Dataset Engineering Team: Built in collaboration with Thareek Anvar M Tools: AI chatbot, custom datasets, Bolt.new ecosystem

BoltHelp was developed as a mentor-driven initiative to support over 100,000 participants in the Bolt.new $1,000,000 AI Hackathon. Acting as a real-time AI assistant, it helped teams and individuals navigate challenges, rules, and technical hurdles by delivering instant, accurate answers—powered by structured internal documentation and official hackathon resources.

As mentors, we noticed recurring questions around prize eligibility, submission guidelines, and technical errors. Time constraints and information overload left many teams confused or stuck, costing them momentum and potential.

We created Bolt Help:

An AI-powered assistant trained on both official hackathon docs, bolt official docs and live mentor-generated documentation

Able to answer participant questions 24/7 during the event

Accessible directly through a lightweight, responsive web interface

Finalized the user interface, ensuring clarity, accessibility, and responsiveness under tight timeframes

Engineered and delivered the curated datasets—including mentor-authored docs and annotated official rules—to enable relevant, accurate bot responses

Collaborated with Thareek Anvar M (who led initial app layout) in a 50/50 co-build partnership

Used by hundreds of thousands of participants to resolve questions instantly

Reduced mentor workload significantly by automating repetitive Q&A

Helped teams and individuals focus more on building and shipping, rather than rule clarification

AI can extend mentorship at scale in time-sensitive environments

Documentation paired with smart delivery (via chatbots) creates massive leverage

Effective UI and data prep are just as critical as underlying model behavior in user-facing tools