Projects
Hands-on architecture implementation guides and reference builds for key AI engineering patterns.
Foundational Builds
3 projectsAPI-to-SQLite CLI Tool
Pull data from a public API on a schedule, store it in SQLite, and serve it through a small FastAPI app.
Basic RAG Chatbot
Upload a PDF, chunk it, embed it, and chat with it. The 'hello world' of RAG — but done properly.
Data Pipeline with Tests
Build a robust local data pipeline that reads messy CSVs, applies an algorithmic transformation, validates data, and writes clean JSON.
Production Architectures
4 projectsEnd-to-End Churn Model
Train a RandomForest classifier on tabular customer churn data, evaluate with precision/recall, export it, and serve predictions.
PyTorch Training Loop from Scratch
Build a neural network from scratch in PyTorch (no high-level wrappers). Write the Dataset, DataLoader, loss function, and training loop.
RAG Evaluation Harness
Build a scoring pipeline that measures retrieval precision/recall and answer faithfulness automatically.
Structured Extraction Pipeline
Take raw messy text (like resumes or receipts), use an LLM API with function calling, and guarantee valid JSON extraction.
Enterprise Systems
5 projectsHybrid Search RAG App
Build a Retrieval-Augmented Generation app using LangChain, Pinecone, hybrid search (keyword + semantic), and a cross-encoder reranker.
Multi-Agent Orchestration System
An orchestrator agent that delegates subtasks to specialized worker agents and merges their outputs.
Multi-Agent Researcher
Use LangGraph or AutoGen to build a research assistant that plans a query, scrapes sources, grades relevance, and writes a cited report.
Transfer Learning Vision API
Fine-tune a pre-trained ResNet model on a custom dataset and deploy it behind a simple Gradio UI or FastAPI endpoint.
Transformer from Scratch
Implement a scaled-down GPT-style decoder-only transformer in PyTorch. Train it on a tiny text corpus to generate characters.