name : Amit Chandra
role : AI/ML Engineer @ Vervebot INC
location : New Delhi, India
education: B.Tech — Computer Science & Engineering (AKTU)
focus : AI Agents · RAG Systems · LLM Fine-tuning · Deep Learning · MLOps · PyPI
passion : Building & training intelligent systems that solve real-world problemsFrom designing neural network architectures and curating datasets to fine-tuning LLMs and shipping production AI agents — I own the full ML lifecycle end to end. My work spans computer vision models, RAG pipelines, autonomous agents, and conversational systems, all built to solve real business problems at scale.
AI-powered recommendation engine published on PyPI — build intelligent recommendation systems with advanced ML algorithms, collaborative filtering, and deep learning. Production-ready, battle-tested.
The FastAPI of AI Agents — a provider-agnostic Python framework for building production-ready AI agents, multi-agent teams, RAG pipelines, and DAG-based workflows across OpenAI, Anthropic, Gemini, Groq & more, with built-in memory, observability, and one-line HTTP deployment.
A unified, local-first AI agent gateway — one assistant across Telegram, Discord, Slack, WhatsApp & 10+ platforms, with a 3-tier memory system (Redis + Qdrant + SQLite), smart LLM routing across Claude, Gemini, DeepSeek, GPT-4 & Ollama, voice support, and a plugin SDK.
| Project | Description | Stack | Status | |
|---|---|---|---|---|
| 🧠 | SchemaMind | AI agent for businesses — connect your database and ask anything in natural language using LLMs + RAG | LangChain OpenAI FastAPI RAG |
 |
| 🔮 | Cortext Flow | Advanced autonomous AI agent — next evolution of OpenClaw. Multi-step reasoning, dynamic tool use, self-improving | GPT-4 LangChain LlamaIndex |
 |
| 🧠 | RecoForge Edge | Browser-native recommendation intelligence. Privacy-first, cross-platform, powered by LLaMA 3.2 3B + RL | Django LLaMA RL Next.js |
 |
| Tripmind | Full-stack AI travel platform with smart itinerary planning, real-time booking & conversational AI | Next.js Django OpenAI PostgreSQL |
|
| Project | Description | Stack | Link | |
|---|---|---|---|---|
| 📊 | Content Recommendation System | Production-grade hybrid recommendation engine. Collaborative filtering + deep learning. Real-time user behavior analysis, <100ms latency | TensorFlow Flask Redis Python |
 |
| 🌾 | AI Crop Recommendation | ML-powered crop selection for farmers. Analyzes soil, climate & regional data for optimal yield predictions | Scikit-learn Pandas Flask Python |
 |
| 🧩 | Sudoku Solver | Upload a Sudoku image or enter manually — CV model extracts digits, scikit-learn handles recognition, custom backtracking algorithm solves it. Downloadable results, error handling, polished UI | Flask OpenCV scikit-learn Pillow JavaScript |
|
| 🎯 | AI Customer Support Agent | Enterprise-grade RAG-powered support agent. Learns from knowledge base & historical tickets to auto-resolve tier-1 queries, analyze customer sentiment, detect churn risk, and escalate to humans when needed. Integrates with Zendesk, Freshdesk & Intercom | RAG LangChain OpenAI FastAPI Vector DB |
 |
| Game | Description | Tech | Play | |
|---|---|---|---|---|
| 🐍 | Snake Game | Classic snake with smooth controls, score tracking & increasing difficulty | JavaScript Canvas API |
 |
| 🧱 | Brick Breaker | Arcade-style brick breaker with physics, power-ups & multiple levels | JavaScript Canvas API |
 |
| 🏓 | Pong Game | Two-player Pong with smooth paddle controls & real-time collision detection | JavaScript Canvas API |
 |
Open to collaborating on AI/ML projects, LLM-based systems, and innovative open-source work. Don't hesitate to reach out!