Open to opportunities

Aayush Dey

@aayushdey

Message

AI/ML and GenAI engineer building production-ready systems across vision, RAG, and LLM orchestration.

India

Message

What I'm looking for

I’m seeking AI/ML Engineer or GenAI Engineer roles where I can contribute from day one—building production-grade pipelines, RAG/LLM orchestration, and reliable MLOps that ship dependable AI features.

I’m a final-year Computer Science engineer focused on building and deploying end-to-end ML systems—from computer vision to GenAI applications and production MLOps.

I’ve delivered 4 real-world projects, including a YOLOv8 plant disease detector with Groq-powered treatment-plan generation, and a production telco churn pipeline with MLflow tracking, Dockerized deployment, and GitHub Actions CI/CD. I also build retrieval-augmented systems—grounding LLM outputs in uploaded documents to reduce hallucinations and improving response quality through vector search.

My work centers on reliable orchestration and evaluation: multi-agent routing with LangChain, fault-tolerant model fallback between Gemini and Groq, and LLM evaluation practices. I’m looking for AI/ML Engineer or GenAI Engineer roles where I can contribute from day one and help ship dependable AI products.

Experience

Work history, roles, and key accomplishments

Telco Churn Prediction

May 2026 - Present (2 months)

Delivered an end-to-end production churn prediction system using an XGBoost + scikit-learn pipeline spanning validation, feature engineering, training, evaluation, and inference. Implemented MLflow tracking, Dockerized the FastAPI (Swagger) + Gradio app, and automated CI/CD with GitHub Actions.

XGBoost scikit learn fastAPI MLFlow Docker GitHub Actions Gradio CI CD Swagger REST APIs

AgriCare AI

May 2026 - Present (2 months)

Architected a full-stack AI agriculture platform combining computer vision, tabular ML, and LLM guidance into a single Streamlit app with cached real-time inference. Built a YOLOv8 plant disease detector and integrated Groq LLM via LangChain to generate treatment plans, plus an XGBoost soil-fertility classifier deployed as a FastAPI endpoint.

PyTorch YOLOv8 OpenCV Lang Chain GROQ Streamlit XGBoost fastAPI REST APIs

IntelliAgent

Nov 2025 - Present (8 months)

Architected a multi-agent LLM orchestration system routing queries across specialized agents for PDF RAG, YouTube Q&A, web research, and an AI resume builder. Implemented fault-tolerant fallback between Gemini and Groq models and built retrieval using ChromaDB with HuggingFace embeddings.

LangGraph Gemini API GROQ ChromaDB Tavily Streamlit Semantic Scholar API Embeddings

AI Document Chat

Aug 2025 - Present (11 months)

Built a full-stack production RAG chatbot that grounds Groq Llama responses in uploaded documents and reduces hallucinations via retrieval over vectorized chunks. Implemented semantic search with HuggingFace embeddings and a ChromaDB-backed store using SQLite, served through an async FastAPI backend with a JavaScript streaming chat UI.

fastAPI ChromaDB GROQ SQLite RAG JavaScript Streaming Responses