AI/ML Engineer — LangGraph, RAG Systems, Multi-Agent Architectures, LLMs
Founder of Neurovia AI — live AI SaaS in production
Built Sebastian — autonomous 59-tool personal assistant with OS control & voice
Open-source author: YouTube Shorts Automation Pipeline
Self-published author — "The Future of Touchless Technology"
CS (AI/ML) · CGPA 8.0 · IDEATHON 2.0 Winner · AWS & GCP Certified
5+ years coding · 30+ projects · 15+ AI/ML models shipped to production
| Project | Description | Stack | Status |
|---|---|---|---|
 Sebastian |
Fully autonomous personal assistant — 59+ tools, OS & browser control, Gemini Live real-time voice, camera/screen vision, 3-tier safety gate (GREEN/YELLOW/RED), multi-layer memory with pgvector, Discord voice bridge, 110+ API integrations | Python FastAPI pgvector Gemini Live Playwright Edge TTS |
🔒 Private |
 Neurovia AI |
AI Social Media Automation SaaS — 4-tier LLM cascade (99.7% reliability), autonomous cross-platform posting, Razorpay billing, OAuth flows, 50+ DAU, 99.5% uptime | Next.js 14 FastAPI PostgreSQL Supabase Gemini OpenAI Docker |
🟢 Live |
| Multi-Agent Data Analytics Platform |
Desktop AI analytics suite — 6-agent LLM pipeline (Profiler→Cleaner→Feature Engineer→Model Selector→Visualizer→Orchestrator), 11 ML models benchmarked, 20+ chart types, offline LLM failover (Groq→Gemini→Ollama) | PySide6 Python Groq Gemini Ollama XGBoost LightGBM |
🟡 GitHub |
 Virtual Therapist |
Stateful 3-agent LangGraph system (intake → emotion-detection → therapeutic-response) with deterministic state transitions, speech-to-text, and adaptive emotional context | LangGraph LangChain Groq ChromaDB FastAPI PostgreSQL |
🟡 GitHub |
| Project | Description | Stack | Status |
|---|---|---|---|
 YouTube Shorts Automation |
End-to-end automated short-form video pipeline: trend scraping (Reddit, HN, Google Trends) → LLM script generation with quality scoring → voice synthesis → FFmpeg video assembly → YouTube upload with metadata. Multi-channel, 4+ videos/run, Windows Task Scheduler for unattended daily posting | Python Ollama FFmpeg edge-tts Pexels API YouTube Data API v3 |
🟢 Open Source |
 Local LinkedIn Automation |
AI-driven LinkedIn outreach automation — profile scraping, personalized connection message generation, automated DM sequences, engagement tracking, and lead pipeline management running fully locally | Python Playwright Gemini API SQLite Scheduler |
🔒 Private |
 Daily Coding Streak Tracker |
Automated daily coding challenge fetcher and submission tracker — pulls problems from LeetCode/HackerRank, logs solutions, and maintains streak stats with GitHub push integration | Python GitHub API Selenium Cron |
🔒 Private |
| Project | Description | Stack | Status |
|---|---|---|---|
| Neurovia AI |
AI Social Media Automation SaaS — LinkedIn & Twitter AI content generation, scheduled posting, 5,000+ users, 99.5% uptime with 4-tier LLM cascade and Razorpay billing | Next.js 14 FastAPI PostgreSQL Supabase Gemini Docker |
🟢 Live |
 Crafting Brain |
Live content-tech platform — NLP pipelines for sentiment scoring, topic extraction & content optimization in production. AWS EC2/S3/Nginx managed infrastructure, GitHub Actions CI/CD | FastAPI Python AWS Docker Nginx |
🟢 Live |
 Ezz Notes |
Real-time PWA note-taking app with offline support, sync, and active user base | React Firebase PWA |
🟢 Live |
 InnerWhispers Blog |
Full-stack blogging platform with social features, JWT auth, rich-text editing, and real-time updates | FastAPI React Supabase JWT |
🟡 GitHub |
 Neurovia Careers |
AI-powered ATS resume builder — weighted scoring engine, 90%+ ATS compatibility across 15+ job categories, 4 custom LaTeX export templates, Groq AI bullet & summary generation | React 18 TypeScript FastAPI Supabase Groq Vite |
🟢 Live |
 Social Media RSS News Feed |
Aggregated multi-source RSS news feed reader with AI summarization, category filtering, and social share integration | Python FastAPI RSS Gemini API React |
🟡 GitHub |
# ═══════════════════════════════════════════════════════════════
# SANGRAMJEET GHOSH · AI Engineer & Full-Stack Developer
# ═══════════════════════════════════════════════════════════════
profile:
education: "B.Tech CS (AI/ML) · Malla Reddy University · GPA 8.0/10"
thesis: "Cost-Constrained AI Automation Systems (Neuroviai) — 15K Python + 12K TypeScript"
location: "Hyderabad, India"
languages: ["English (Professional)", "Hindi (Native)", "Bengali (Native)"]
experience:
- role: "Freelance AI & Full-Stack Developer"
period: "Aug 2025 → Present"
highlights:
- "Delivered end-to-end MVPs for 3+ startup founders"
- "SaaS products with Razorpay billing, JWT/OAuth, multi-platform API integrations"
- "Deployed across AWS · Render · Vercel · Netlify with custom domains & SSL"
- "LLM cost control: quota splitting, response caching, request batching"
- role: "Data Science Intern — Inikola Technologies (CraftingBrain.com)"
period: "Mar 2025 → Dec 2025"
highlights:
- "Built NLP pipelines for sentiment scoring, topic extraction & content optimization"
- "Managed AWS EC2 · S3 · IAM · Nginx reverse-proxy in production"
- "Multi-stage Docker builds · GitHub Actions CI/CD zero-downtime deployments"
skills:
languages: ["Python", "TypeScript", "JavaScript", "SQL", "HTML5/CSS3", "Bash"]
ai_ml: ["LangChain", "LangGraph", "RAG Pipelines", "Multi-Agent Systems",
"Gemini API", "OpenAI", "Groq", "pgvector", "Prompt Engineering",
"XGBoost", "LightGBM", "Scikit-learn", "Pandas", "NumPy"]
backend: ["FastAPI", "REST APIs", "WebSockets", "JWT/OAuth 2.0",
"Microservices", "Async Python", "SQLAlchemy ORM"]
frontend: ["React 18", "Next.js 14 App Router", "Tailwind CSS",
"Zustand", "ShadCN/UI", "Vite", "TypeScript"]
databases: ["PostgreSQL", "Supabase", "pgvector", "Redis", "MongoDB"]
devops_cloud: ["AWS (EC2·S3·IAM)", "Docker", "Docker Compose",
"GitHub Actions CI/CD", "Nginx", "Render", "Vercel", "Netlify", "FFmpeg"]
tools: ["Git", "Postman", "VS Code", "Playwright", "PyAutoGUI"]
currently_building:
name: "Sebastian — Fully Autonomous AI Personal Assistant"
tools: "59+ tools · 110+ API integrations"
highlights: ["OS & browser control", "Gemini Live real-time voice (bidirectional)",
"Camera + screen vision", "3-tier permission gate (GREEN/YELLOW/RED)",
"Discord voice bridge (48kHz→16kHz PCM resampling)",
"Multi-layer memory: episodic + semantic + behavioral (pgvector)"]
I love connecting with different people so if you want to say hi, I'll be happy to meet you! 😊
