Available for full-time roles

DamewanBareh

Backend & AI Engineer · Final-year MCA

I build production backend systems and local-first AI tools — from a BGV module live at an AI recruitment company to a fully offline desktop AI.

GitHub Resume

Production systems, not just side projects.

3 internships · FastAPI, WebSockets, Redis, PostgreSQL, MongoDB

Backend Developer Intern

Aimploy · AI-Powered Recruitment Platform

Apr 2026 – PresentRemote

10+ API endpoints6+ doc types handledLive in production

Designed and shipped a Background Verification (BGV) module to production using FastAPI & PostgreSQL with 10+ REST API endpoints and 5 normalized DB tables covering document upload, AI validation, and employment verification.
Built a dual-storage layer (Supabase + AWS S3) for 6+ document types (Aadhaar, PAN, passport, payslips, etc.); used OpenAI GPT-4 Vision to score documents 0–100, auto-verifying those scoring ≥ 80.
Automated HR employment verification with 7-day expiring tokenized links and majority-rule aggregation across 3+ employers per candidate — the full module is live in production, managed via Git & GitHub.

FastAPIPostgreSQLSupabaseAWS S3GPT-4 VisionREST APIs

Full Stack Developer Intern

Axentra OS · AltissAdvance Tech Pvt. Ltd.

Mar 2026 – Jun 2026Remote · Mumbai, Maharashtra

< 100ms latency500+ concurrent WSsub-ms query on 10k+ msgs

Built a real-time Chat Module using FastAPI WebSockets and React with sub-100ms delivery latency across 4 API endpoints; managed all branches and PRs via Git & GitHub.
Engineered a Redis Pub/Sub broadcasting layer with a custom ConnectionManager supporting 500+ concurrent WebSocket connections across multiple server nodes.
Designed the MongoDB data layer with 5 compound indexes and an async DAO pattern, achieving sub-millisecond query performance on paginated chat histories of 10,000+ messages.

FastAPIWebSocketsReactRedis Pub/SubMongoDB

Web Development Intern

Meghalaya Technology Parks Society

Jun 2025 – Jul 2025Shillong, Meghalaya

25% bounce rate reductionLive dashboard

Redesigned and rebuilt the Meghalaya IT Parks website using HTML, CSS, JavaScript, and PHP; cut page bounce rate by 25% through improved information hierarchy and load-time optimisation.
Developed an FAQ chatbot using rule-based NLP and a live project-tracking dashboard displaying real-time government IT initiative data, reducing manual status enquiries for the team.

HTMLCSSJavaScriptPHPNLP

02 / projects

Built offline-first and production-ready.

All on GitHub

Luna OS

Local AI Personal Assistant

Architected a fully offline, local-first AI desktop application (Electron + React + FastAPI) with real-time LLM response streaming over WebSockets, an agentic tool-calling engine across 10+ capabilities (file I/O, calendar, reminders, contacts, IoT), and a per-invocation permission-gating system that intercepts every tool call before execution.

Fully offline / local-first
10+ agentic tool capabilities
Per-invocation permission gating
Windows installer via PyInstaller + electron-builder

PythonFastAPIReactElectronOllamaSQLiteFaster-WhisperWebSockets

NexNote

AI Study Assistant

Built a RAG-based study assistant with a Pinecone vector DB and self-hosted Ollama LLM; implemented semantic search across ingested PDFs, DOCX, and TXT files using nomic-embed-text embeddings.