Resume.PDF

RUDRAKSHBHARDWAJ

Applied AI Engineer · ML · LLMs · Backends

Learning by doing

I'm an Applied AI Engineer building production-grade AI systems— multi-model orchestration, RAG pipelines, and distributed backends — for engineering teams shipping real products.

Currently a Summer Intern at EY, building internal treasury tooling and AI-assisted workflow automation.

WritingsTechnical & non-technical essays.
Rudraksh Bhardwaj

Featured Projects

LLM-Consulate

Multi-model AI consensus platform

Consults a council of open-source LLMs in parallel, measures inter-model agreement, preserves dissenting viewpoints, and synthesizes answers only when meaningful consensus exists.

  • Parallel multi-model orchestration with real-time SSE streaming
  • Consensus scoring with dissent preservation when agreement is weak
  • Production FastAPI backend with async inference coordination
FastAPINext.jsTypeScriptSSENVIDIA AI

YouTubeGPT

Video-grounded RAG for YouTube Q&A

YouTubeGPT

Video-grounded RAG for YouTube Q&A

Full-stack GenAI application that ingests YouTube transcripts, indexes dense embeddings in ChromaDB, and serves context-aware answers through a REST API with structured evidence display.

  • 30% answer relevance improvement over naive retrieval baselines
  • Decoupled retrieval and generation for independent optimization
  • Collapsible Q&A history with semantic transcript chunking
ReactFlaskLangChainChromaDBNVIDIA AI

GitGallery

GitHub-backed photo vault

GitGallery

GitHub-backed photo vault

Desktop application that repurposes GitHub repositories as a photo storage backend with OAuth authentication, thumbnail generation, multithreaded synchronization, and folder organization.

  • Background worker threads for non-blocking Git and sync operations
  • Full GitHub OAuth flow with repository lifecycle management
  • Local cache mapping folder hierarchies to Git tree structures
PythonPySide6GitHub OAuthPillowMultithreading

Skills

AI & ML

LLM SystemsRAGLangChainLangGraphPyTorchYOLOv8Computer VisionAgentic Workflows

Backend & Systems

FastAPIFlaskPythonNode.jsSSE StreamingAsync OrchestrationREST APIs

Frontend

ReactNext.jsTypeScriptTailwind CSS

Data & Infrastructure

ChromaDBPostgreSQLMongoDBVector DatabasesGit

Education

  1. 2023 - Present

    B.Tech in Computer and Communication Engineering

    @ Manipal University Jaipur


Other Projects

  1. RoadDefect-Dodge

    YOLOv8 road defect detector trained from scratch

    End-to-end computer vision pipeline for pothole, crack, and manhole detection — dataset curation, bounding box annotation, augmentation, training, validation, and inference with no pretrained weights.