Skip to content
View igennova's full-sized avatar

Block or report igennova

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
igennova/README.md

Lalit Negi

Backend engineer. I build real-time systems, AI pipelines, and open-source infrastructure.

Experience

Google Summer of Code 2025 — OWASP Foundation
Built a WebRTC video engine, live leaderboard with Redis and WebSockets, and a CTF-style security simulation lab.

AI Engineer — PUCH AI
Built a multi-tool LLM agent with Redis + Lua rate limiting stable for 10k+ users; unified image/video generation into a single inference layer.

Backend Developer — Persist Venture
Architected an AI-powered media pipeline using FastAPI and PostgreSQL.

Projects

Distributed Vector Search Engine — ANN search from scratch
HNSW nearest-neighbor search built from scratch (0.93 recall@10 at 2.5× lower latency vs exact). Sharded across gRPC services with a scatter-gather coordinator; cut 4-shard latency 3.5× (5.6 → 1.6 ms) after profiling GIL contention. Python + gRPC + NumPy.

WeMakeVideos — AI text-to-video SaaS (demo landing)
End-to-end prompt → script → voice → scenes → render workflow built with Next.js + TypeScript + Remotion. Public demo showcases the landing + product preview; full app walkthrough available on request.

ZeroCostShorts — Automated AI video pipeline (76★)
Fully autonomous pipeline (LLM → TTS → image gen → FFmpeg render → YouTube upload) running daily via GitHub Actions at ~$0/day. Forked by others for their own workflows.

LeetFight
Real-time 1v1 coding arena. Players compete on the same problem simultaneously with live code sync, instant judging, and an ELO rating system. 180+ active users, sub-50ms WebSocket latency.

Open Source Contributions

  • @PostHog ( Link )
  • @antiwork ( Link )
  • @antiwork/gumboard ( Link )
  • @antiwork (bounty PRs via @devwill2) ( Link )
  • @OWASP-BLT ( Link )
  • @openml ( Link )
  • @getnao ( Link )
  • @PalisadoesFoundation ( Link )

Fun Projects

text-to-video-engine — AI video generation
seedhe-maut-player — Custom music player
anime-predictor — LeetCode to anime matcher
doc-o-drop — AI health and wellness platform

Stack

  • Backend & APIs: FastAPI, Node.js, Django, REST, WebSockets, WebRTC, gRPC
  • Frontend: Next.js
  • Databases & Caching: PostgreSQL, Redis (with Lua scripting)
  • Queue & Jobs: BullMQ
  • DevOps & Infra: Docker, AWS (Lambda), CI/CD
  • Languages: TypeScript, Python, C++
  • Concepts: System Design, Scalable Systems, Microservices
  • Other: Remotion, FFmpeg, Supabase

Highlights

  • $20,000+ in open-source bounties
  • GSoC 2025 contributor at OWASP
  • Multiple hackathons won
  • LeetCode 1815

Open to collaborating on backend systems, real-time infrastructure, and AI pipelines.

luckynegi1025@gmail.com · Portfolio

Pinned Loading

  1. ZeroCost-Shorts ZeroCost-Shorts Public

    Fully automated faceless YouTube Shorts machine, running on autopilot. AI writes the script, AI speaks it, AI generates the visuals, FFmpeg stitches it all together, and it uploads to YouTube daily…

    Python 74 22

  2. PostHog/posthog PostHog/posthog Public

    🦔 PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture…

    Python 40k 3.4k

  3. openml/server-api openml/server-api Public

    Python-based server

    Python 17 50

  4. Distributed-Vector-Search-Engine Distributed-Vector-Search-Engine Public

    Distributed vector search engine built from scratch in Python: HNSW indexing, sharding across nodes, replication, failure handling, and recall/latency benchmarks.

    Python 1