I am
Aditya Dey.
Backend and AI/ML Engineer with 2.5+ years building production RAG systems, intelligent document processing pipelines, and agentic AI on AWS. AWS-certified ML Engineer — ships end-to-end systems that automate complex, high-volume workflows.
ToolsoftheTrade
Technologies I use to build fast, reliable, and scalable systems.
Gen AI & Agents
ML & MLOps / LLMOps
Backend & APIs
Cloud & DevOps
Databases & Vectors
WorkHistory
Dec 2023 – Present
Full-timeBackend Developer - AI
Workmates Core2Cloud
Building production AI systems at scale — 2M+ document intelligence pipeline with AWS Textract & Qdrant RAG, multi-template invoice extraction with Bedrock LLM normalisation, sub-100ms policy recommendation via pgvector hybrid search, WhatsApp NBFC loan advisory chatbot, and an autonomous CloudWatch incident-response agent with LangGraph. Leads GenAI development and mentors junior engineers.
Aug – Oct 2023
InternshipML Engineer Intern
Prodigy Infotech
Built ML models for NLP and recommendation systems using scikit-learn and PyTorch. Contributed to production pipelines and improved user experience through data-driven feature work.
WhatIBuild
End-to-end AI engineering — from model experimentation to production deployment.
Gen AI & Agentic Systems
Multi-agent LangGraph workflows, RAG pipelines, tool-use agents, and context-aware chatbots for production environments.
Voice AI & Real-time
End-to-end voice AI pipelines with LiveKit, real-time STT/TTS, and streaming multimodal interfaces.
ML Engineering
Train, evaluate, and ship ML models for NLP, classification, and recommendation — from notebook to API.
MLOps & LLMOps
Full model lifecycle management — monitoring, evaluation, versioning, observability, and automated deployment pipelines.
Backend & APIs
Production-grade FastAPI services, event-driven microservices with Kafka, and scalable Python backends.
Cloud & Deployment
AWS infrastructure, containerised deployments on ECS, CI/CD pipelines, and IaC with Terraform.
HowIWork
Understand
I start by deeply understanding the goal and the problem space before writing a line of code.
Prototype
I build quick prototypes to test assumptions and gather early feedback fast.
Iterate
I refine the solution based on real-world usage, performance metrics, and edge cases.
Deliver
I ship clean, documented, production-ready code with proper testing and observability.
Let's Build Something Great
Have a project, idea, or just want to talk AI? Let's make something great.