Projects
Here you'll find a collection of my recent projects demonstrating my expertise in production AI engineering, cloud infrastructure, and DevOps. Each project reflects a focus on building systems that are not just functional but reliably deployable and maintainable at scale.
Featured Project Details
AI Chatbot Platform on AWS EKS
A production-grade, full-stack AI chatbot deployed on AWS EKS with a three-tier architecture (Next.js frontend, FastAPI backend, PostgreSQL database). Inter-service traffic stays inside the cluster via ClusterIP, with only the frontend exposed externally.
Technical Highlights:
- Jenkins CI/CD pipeline that version-bumps, builds and pushes tagged images to ECR, and performs a zero-downtime Helm rollout to EKS on every git push
- LangGraph AI agent with dual-layer PostgreSQL memory: short-term per-conversation and long-term cross-session
- Real-time web search tool and personalised user memory read/write integrated into the agent
- Kubernetes-native deployment with Helm charts for repeatable, environment-consistent releases
Agentic RAG Application with Docling, LangChain & PostgreSQL
An agentic RAG system capable of ingesting and querying almost any document format, with embeddings stored in a PostgreSQL database using the pgvector extension.
Technical Highlights:
- Document parsing for PDFs, Word docs, spreadsheets, and more using Docling
- Embeddings stored in PostgreSQL with the pgvector extension for efficient similarity search
- RAG workflow orchestrated and agent built end-to-end with the LangChain framework
- Modular pipeline design enabling easy swapping of models and retrieval strategies
Stock Analyst Deep Agent
A multi-agent stock analysis system with specialised subagents for fundamental, technical, and risk analysis, backed by live data integrations and a FastAPI backend.
Technical Highlights:
- Specialised subagents for fundamental analysis, technical analysis, and risk evaluation
- Yahoo Finance integration for live stock price and financial statement data
- Tavily integration for real-time news and market trend retrieval
- FastAPI backend exposing clean API endpoints for agent orchestration
Automated CI/CD Pipeline with Jenkins and AWS EC2
A fully automated delivery pipeline for a Node.js/Express application, containerised with Docker and deployed to AWS EC2 on every push to the main branch.
Technical Highlights:
- Jenkins pipeline with automatic semantic versioning, test execution, Docker image build and push to Docker Hub
- Deployment to AWS EC2 via SSH with Docker Compose for zero-touch releases
- Jenkins shared library managing the full delivery lifecycle for reusability across projects
- Docker & Docker Compose containerisation ensuring environment parity from dev to production
Interested in collaboration?
Let's discuss how we can work together on innovative AI solutions.