Built a production Agentic AI microservice from scratch as the primary engineer at Techkhana, serving real clients over 360 working hours.
Key features built:
- Full service architecture designed and owned end-to-end
- FastAPI backend with clean layered architecture
- AWS Bedrock integration for cloud-based LLM capabilities
- Multi-agent orchestration pipelines using LangGraph
- RAG (Retrieval-Augmented Generation) systems with LangChain
- PostgreSQL for data persistence
- Deployed on AWS infrastructure
- Resolved real production challenges: performance bottlenecks, model inconsistencies
- Also contributed to Construction Management Platform backend (Node.js + PostgreSQL)
Officially certified by Techkhana CTO — 360 production hours.