Back to Home

Engineering Log

A selection of complex systems I have architected.
Focusing on AI Agents, RAG Pipelines, and Developer Tools.

CASE STUDY Automated Workflows

FinTech Agent Bot

Enterprise Financial AI Agent

The Challenge

Static LLMs hallucinate on real-time financial data and cannot execute secure banking transactions.

The Solution

Engineered an agentic workflow using Google ADK to route queries. The agent intelligently selects tools (Stock API, Transaction Engine) based on user intent.

System Architecture

User
Google ADK
Gemini API
Fin. Tools

Technology Stack

PythonGoogle ADKGemini API
CASE STUDY Reduced Support Tix

Aaruush Virtual Assistant

RAG-Powered Event Support

The Challenge

Tech fest attendees struggled to find specific event details buried in massive PDF schedules and websites.

The Solution

Built a Retrieval-Augmented Generation (RAG) pipeline. It embeds event docs into ChromaDB, allowing the LLM to answer context-aware queries instantly.

System Architecture

React UI
FastAPI
ChromaDB
LLM Inference

Technology Stack

LangChainChromaDBFastAPIReact

More Experiments