Triple Hackathon Winner

Three projects, one platform: A developer's winning streak with Fireworks AI

GLM 5.2 is live! Opus-level intelligence at open-source rates. Pay per token on serverless. Try it today.

Blog

Triple Hackathon Winner Three projects, one platform: A developer's winning streak with Fireworks AI

PUBLISHED 10/14/2024

Table of Contents Core architecture Technical implementation From preparation to execution

Pre-hackathon preparation

During the hackathon Nehil Jain: Unstoppable builder and AI Engineer Selvam Palanimalai: Technical architecture expert A track record of excellence

Hackathon 1: KinConnect - Optimizing hackathon team formation

Technical stack

Hackathon 2: LazyPMs Inc - Automating release notes with multi-agent AI

System architecture Words of wisdom

Table of Contents

When it comes to building with Fireworks AI, few developers can match Nehil Jain's track record. His latest triumph – securing second place at the E2B x Fireworks AI hackathon alongside Selvam Palanimalai – marks his third consecutive success using the platform. From automating release notes with LazyPMs to matching hackathon teams with KinConnect, and now ensuring documentation reliability with ProoferX, Nehil has consistently demonstrated the power and versatility of Fireworks AI. "The speed I get using Fireworks endpoints for Llama models is one of the key drivers for successful outcomes," explains Nehil, whose deep understanding of the platform has been crucial to his winning streak. "A combination of Llama for intelligence and Firefunction for tool calling, along with structured outputs, lets me build reliable AI pipelines." His latest project, ProoferX, developed with technical architecture expert Selvam Palanimalai, tackles one of the most persistent challenges in the developer tools ecosystem: maintaining reliable technical documentation. You can see their solution in action in their demo on X or check out all the hackathon demos on YouTube . The problem: documentation in crisis

"Good tech products need solid docs and examples. Developer experience and trust are essential for these products to grow," explains Nehil Jain, co-creator of ProoferX. "Out-of-date code examples in public technical documentation can really hurt the product builders." The team uncovered several real-world examples that validated their concerns: • OpenAI's structured output introduction contained broken references to data files • Vite's installation docs had issues that could frustrate junior developers • E2B's Python API key documentation contained a small but significant typo

The challenge, Nehil points out, goes beyond simple maintenance: "It is unimaginable to expect a team to manually check that their guides work for all the versions of the libraries used in the examples for all the OSes for different versions of compilers the user is using. The only way to achieve this is through automation."

From E2B’s Tereza Tizkova LinkedIn post, “🏆 Who won our first hackathon?“ The solution: ProoferX

Core architecture

Drawing on his experience with Fireworks AI, Nehil worked with Selvam to architect a solution that could tackle this complex challenge. ProoferX operates through three crucial steps: Code Extraction : Analyzes blog posts and documentation to create complete, end-to-end working code snippets Success Definition : Determines the intended outcome for readers Sandbox Validation : Executes code in a fresh environment to replicate the reader's experience

Technical implementation

At the heart of ProoferX lies a carefully orchestrated pipeline that begins with Firecrawl converting documentation URLs into markdown format.

*Diagram from Nehil Jain’s blog post, “ How we won Code Interpreter 2.0 with ProoferX ”.* Code extraction pipeline • Uses Firecrawl to convert URLs into markdown • Employs Llama 3.2 models via Fireworks AI for information extraction • Implements Pydantic for structured data handling and validation

Analysis flow URL → Markdown → Code Snippets → Success Criteria → Validation Key components • crewAI application for iterative code extraction • E2B sandboxes for code execution • Structured output with Pydantic models • Python dependency management system

Agentic approach • Moved beyond simple Chain of Thought prompting • Implemented iterative extraction for better recall • Handled multiple code illustrations in single documents • Used E2B Code Sandbox for code execution

Check out Nehil’s blog for a full video demo of ProoferX and more details about his and Selvam’s personal perspective on their most recent hackathon. From preparation to execution

The team's success wasn't just about technical innovation – it was built on methodical preparation and execution. Pre-hackathon preparation

• Researched sponsor guides and available credits • Pre-formed team with complementary skills • Prepared starter kits and relevant documentation • Used Cursor IDE for efficient development

During the hackathon

During the event, their process was equally structured. • Created a clear data flow diagram • Implemented 45-minute sync cycles for code merging • Started with simple prompt chains to understand LLM capabilities • Used Firecrawl for URL-to-markdown conversion • Implemented Pydantic for structured data handling

The team behind ProoferX

Nehil Jain: Unstoppable builder and AI Engineer

Nehil brings deep expertise as: • Founder of DemoDrive • Former engineer at QuantumBlack, McKinsey, and Super.com • Expert in AI engineering, focusing on GenAI, MLOps, and modern data platforms • Track record of transforming complex business challenges into practical, scalable AI solutions

"We have been building data platforms, infra, and engineering pipelines for various use cases for a long time in our careers," Nehil explains. "Using ML and now AI on top of it is a natural progression to unlock more value." Selvam Palanimalai: Technical architecture expert

As co-founder at DemoDrive, Selvam (ex-Zomato, Rubikloud, and Kinaxis) contributed crucial expertise: • Based in Toronto, traveled to SF specifically for this hackathon • Expert in Cloud infrastructure, Application development • Focused on building tools to help devtools scale their communities and user bases

The team's complementary skills proved crucial for ProoferX's success, with Nehil handling LLMs and prompting while Selvam managed code sandboxes and app scaffolding. Technical mastery with Fireworks...