Headline

Full Stack Developer @ Dimaag AI | React · TypeScript · FastAPI · Python · AWS | Building real-time AI & agentic systems (RAG, MCP, LangGraph)

About

Full Stack Developer with ~4 years building scalable, real-time products end to end. At Dimaag AI I ship across the stack — React/TypeScript on the front, FastAPI/Python, PostgreSQL, and Redis on the back — with a focus on real-time data systems (WebRTC, SSE), AI-integrated backends (RAG, MCP, LangChain/LangGraph), and high-performance frontend architecture. Recent work I'm proud of: - A live video inference pipeline synchronizing model outputs with WebRTC streams — cutting insight delivery from hours to seconds. - A schema-driven, configurable onboarding system that dropped client onboarding from months to 2–3 days. - An automated AI code-review service (GitHub CLI + Claude) monitoring 6 repos, turning review turnaround from a full day into instant feedback. - A 25–30 component internal UI library (shadcn/ui, Radix, Tailwind) standardizing frontend work across the team. I like turning messy, data-heavy problems into clean, fast, production systems. Previously moved from frontend-only React into full-stack, and before tech I trained as a civil engineer and designer — so I bring a builder's eye to how things fit together. Open to SDE-2 / mid-level full-stack and agentic-AI roles.

Location

Shivamogga

AI stack

anthropic apis

claude code

claude

gemini

langchain

mcp

Work experience

Full Stack Engineer at DIMAAG-AI, Jan 2023 - Present · 3 yrs 10 mos, Bengaluru, - Redesigned the results dashboard into a schema-driven, configurable system supporting multiple AI workflows — reducing client onboarding from months to 2–3 days and contributing to closing 2 enterprise clients. - Engineered the company's first live video inference pipeline, synchronizing model outputs with WebRTC streams in real time — cutting insight delivery from hours to seconds. - Architected message-driven distributed systems using AWS SQS/SNS for scalable microservice communication and real-time AI processing. - Implemented backend-to-frontend SSE streaming with incremental payload delivery, enabling smooth rendering of 1–2M time-series data points via uPlot virtualization. - Built an automated AI code-review service (GitHub CLI + Claude) monitoring 6 repos, cutting review turnaround from a full day to instant feedback. - Built and maintained a 25–30 component internal UI library (shadcn/ui, Radix UI, Tailwind), standardizing frontend development across a 4-developer team. - Prototyped an Observability MCP server unifying the observability stack (SigNoz, Prometheus), enabling trace/span-ID-based root-cause lookups across logs, metrics, and traces to speed up debugging of failed API calls.

3D Designer at Nolte India, Mar 2021 - Mar 2022 · 1 yr 1 mo, Bengaluru, Karnataka, India

Education

civil engineering, Jawaharlal Nehru national college of engineering

Public links

github: https://github.com/vivek4012