Sudharsan Jain

Software Engineer

Open to Work

About

Software Engineer specializing in AI systems, backend architecture, and full-stack development. Experienced with Rust, Go, TypeScript, and Python. Building scalable systems for LLM applications and AI Twins. Contributing to open source projects and maintaining developer communities.

Experience

Founding Software Engineer at myOrbit
myOrbit | Remote | Sep 2024 – Jan 2026
  • Architected multi-layer memory system for AI Twins with Redis caching and conversation summarization; built RAG pipeline using pgvector and OpenAI embeddings
  • Developed "Spine" prompt orchestration system with modular cords for context injection, memory retrieval, and personality adaptation
  • Implemented multi-agent AI architecture in Go with specialized agents coordinated through an orchestrator with streaming responses
  • Led migration from Python to microservices: Bun Elysia API, Socket.IO with Redis Streams, Go workers with Uber FX
  • Implemented Infrastructure as Code using SST.dev for AWS (ECS Fargate, Lambda, RDS PostgreSQL + pgvector, ElastiCache)
Open-Source Maintainer at Keizerworks
Keizerworks | Remote | Oct 2024 – Jun 2025
  • Maintained multiple open source applications, contributing to community growth and establishing best practices
Founding Software Engineer at zeed
zeed | Chennai, India | Oct 2023 – Jul 2024
  • Built referral and rewards program that fueled user acquisition and retention
  • Developed HUID verification flow for jewellery quality assurance
Founding Software Engineer
Zap | Bangalore, India | Jun 2023 – Sep 2023
  • Built and published SDK using rollup; implemented IaC using SST.dev and Turborepo
Founding Software Engineer at Qwk
Qwk, The Convenience App | Chennai, India | Apr 2022 – Jun 2023
  • Sole Full-Stack Developer: built responsive UI with React Native and backend with Express.js, GraphQL, PostgreSQL

Projects

ctxe
In Progress | github.com/Sudarsh1010/ctxe
  • Building context optimization tool for AI coding agents (Claude Code, OpenCode, Cursor) using AST-based compression and semantic search
  • Implementing tree-sitter based code compression achieving 60-70% token reduction by extracting function signatures, type definitions, and module interfaces
  • Designing proactive token budget enforcement and semantic code search with embeddings to replace primitive truncation strategies

Skills

I work across languages and frameworks. I pick tools based on constraints, not trends.

Languages

JavaScript, TypeScript, Go, Python, Rust, SQL

Frameworks

Next.js, React Native, Express.js, Hono.js, tRPC, Django, Node.js, HTMX, A-H Templ

Tools

Docker, Git, PostgreSQL, SQLite, Redis

Platforms

AWS, Cloudflare, Vercel, Supabase