Retrieval and multi-tenancy2025
Multi-Tenant RAG API
A document-QA backend where every tenant's data is walled off from every other tenant's — the first question any firm asks before handing over their files.
FastAPIChromaDBLangChainGeminiJWT
The lab
What I build when nobody's paying me.
Open-source builds where I work out the hard parts — tenant isolation, retrieval quality, reproducible training — before they show up in client work. Every one is public; the code is the claim.
Reproducible ML pipelines2026
YouTube Sentiment Insights
An end-to-end sentiment classifier built the way production ML has to be built — versioned data, tracked experiments, a registered model, and an API at the end of it.
DVCMLflowXGBoostscikit-learnFlask
Email automation2024
Bulk Mail Sender Bot
A scripted, templated email dispatcher — the earliest ancestor of the inbox automation I now build for firms.
PythonSMTPAutomation
Time-series deep learning2024
Climate Forecasting with LSTM
A recurrent sequence model for climate time series — the problem shape behind any forecast that has to respect what came before it.
TensorFlowLSTMTime series
Regression2024
Flight Price Prediction
A tree-ensemble regressor on messy, real-world pricing features — mostly an exercise in feature engineering, which is where the accuracy actually lives.
scikit-learnRandom ForestFeature engineering
These are my own builds, not client engagements — so there are no performance numbers attached to them. For work that ran in production and has the results to show, see the case studies.
Losing hours to work an AI system could handle?
Tell me about the process. I'll tell you straight whether it's worth automating — no pitch, no obligation.
Book a call