Open to opportunities

Sourav Dey

@souravdey3

Message

Full-stack engineer building production GenAI apps with Python, React, and RAG—scalable, cloud-native, CI/CD-driven.

India

Message

What I'm looking for

I want to build production LLM/RAG products end-to-end—FastAPI and React frontends, cloud-native microservices on GCP, and strong CI/CD—so we can improve retrieval accuracy, agent reliability, and ship fast with real business impact.

I’m a Full Stack Engineer (Python + React) and Generative AI specialist with 2+ years building production-grade LLM-powered applications and RAG architectures. I focus on shipping end-to-end systems that deliver measurable improvements in retrieval quality, reliability, and speed.

At Deutsche Bank, I built an end-to-end RAG-powered enterprise search across JIRA/Confluence using LangChain and ChromaDB, deployed on GCP Vertex AI—improving retrieval efficiency by 80%. I also developed compliance AI agents with contextual embeddings (pgvector similarity search), reducing false positives by 30% and accelerating alert resolution, and built FastAPI backends with React/TypeScript frontends on GCP Cloud Run.

As Founder & Lead Full Stack AI Engineer at Cognition Catalysts, I designed and shipped 5+ LLM-powered applications using Python and FastAPI with React/TypeScript, integrating Pinecone and FAISS for semantic retrieval. I’ve built scalable RAG pipelines with LlamaIndex and LangChain, applied prompt engineering for behavior and response quality, and developed GCP-based AI accelerators with full CI/CD automation—plus led a small team of engineers on an applied healthcare LLM product.

I’m passionate about building AI accelerators, MVPs, and innovation-led solutions that turn models into dependable products. I enjoy owning cloud-native microservices, data pipelines, and automation so teams can iterate faster and operate confidently.

Experience

Work history, roles, and key accomplishments

Current

Founder & Lead Full Stack AI Engineer

Current

Cognition Catalysts

Nov 2024 - Present (1 year 8 months)

Founded and led delivery of 5+ LLM-powered applications, building RAG pipelines with LlamaIndex and LangChain and integrating Pinecone and FAISS for semantic retrieval. Built GCP Cloud Run/Cloud Functions AI accelerators and FastAPI services, including a Gemini Vertex AI healthcare prescription parsing system.

Python fastAPI React TypeScript LlamaIndex LLMs (LangChain RAG Pinecone Faiss Prompt Engineering CoT Cloud Run Cloud Functions Gemini Vertex AI Text To BPMN REST APIs

Current

Software Development Engineer I

Current

Deutsche Bank

Jul 2024 - Present (2 years)

Built a RAG-powered enterprise search using LangChain/LangGraph with ChromaDB and deployed it on GCP Vertex AI. Developed React + FastAPI GenAI features on Cloud Run, and delivered cloud-native Python/Java microservices with Docker/Kubernetes and CI/CD using GitHub Actions and Cloud Functions.

Lang Chain LangGraph RAG ChromaDB Vertex AI React fastAPI Cloud Run AI Agents Pgvector Docker Kubernetes Helm GitHub Actions Cloud Functions Microservices Spring Boot PostgreSQL Apache Kafka

Software Engineering Intern – ML

Alphagrep Securities

Feb 2024 - Jul 2024 (5 months)

Built a Python-based LLM-integrated Slack bot to automate JIRA management and scheduling. Implemented real-time ML anomaly detection pipelines for trading data using Isolation Forest, LOF, and KNN.

Python LLM Slack Automations Jira Isolation Forest K NN Machine Learning

Software Engineering Intern – Backend

Deutsche Bank

Jun 2023 - Aug 2023 (2 months)

Implemented 15+ REST APIs using Spring Boot with OAuth 2.0 SSO for 500+ concurrent users. Improved code quality with JUnit5/Mockito, reaching 95% test coverage and reducing production bugs.

Spring Boot REST APIs OAuth 2.0 Java Mockito