I build low-latency, production-minded systems — real-time collaboration, Docker-sandboxed code execution, and RAG pipelines that stay predictable when data grows.
Highlights
- Apache Airflow — merged PRs into widely used production OSS
- DevCollab — multi-user coding with ~1–5ms sync and sandboxed runs
- Codebase RAG — 4,300+ chunks indexed, ~11ms retrieval via FAISS
- Ask me about WebSockets · Docker · RAG · Airflow
const lohith = {
location : "Hassan, Karnataka 🇮🇳",
focus : ["real-time systems", "OSS", "RAG"],
mantra : "Ship, measure, harden, repeat.",
};flowchart LR
subgraph RT["Real-time"]
A[WebSockets]
B[Sandboxed runtimes]
end
subgraph AI["Data & AI"]
C[RAG]
D[Vector search]
end
subgraph Ops["Ship"]
E[Docker]
F[CI / Git]
end
RT --> Ops
AI --> Ops
Languages
Frontend
Backend & real-time
Data, AI & search
DevOps
Real-time collaborative coding — Docker-sandboxed, built for speed.
Multi-user sync with sandboxed execution so nobody nukes the host.
10+ concurrent users · 1–5ms sync latency · ~1.3s sandbox execution
Ask your codebase anything — answers in milliseconds.
Chunk, embed, retrieve with FAISS + LangChain. Works at real scale.
4,300+ code chunks indexed · ~11ms query latency (FAISS) · 100% test pass rate
Images with auth and CDN delivery that holds up.
Google OAuth, JWT, Cloudinary — wired properly end-to-end.
Google OAuth + JWT auth · Cloudinary CDN delivery · Tailwind responsive UI
OSS used at serious scale — real reviews, real merges.
50,000+ GitHub stars · Multiple merged PRs · UI · Docs · Import fixes


