Devin Thakker

iOS and web app that condenses local municipal meeting minutes into digestible daily summaries. Residents can quickly review agendas, decisions, and follow-up actions so they stay engaged with what is happening in their community.

Public-facing RAG web app over Chicago FY2026 budget ordinances with source-grounded answers, page-level citations, and an embedded PDF viewer for opening exact document pages from each result.

Lightweight Python toolkit for calculating live technical indicators across equities and digital assets. Designed to plug into data pipelines so traders can surface moving averages, oscillators, and other analytics without leaving their workflow.

Backtesting framework for quantitative strategies that helps researchers simulate trades against historical price series. Provides utilities for configuring entry and exit signals, executing parameter sweeps, and reviewing performance metrics side by side.

PyTorch exploration into modeling housing prices with neural networks. Includes data preprocessing pipelines, feature engineering experiments, and training notebooks to benchmark architectures and evaluate prediction accuracy.

Machine learning project focused on surfacing anomalous activity in credit card transactions. Applies data balancing techniques and multiple classification models to improve recall on rare fraudulent events while keeping false positives manageable.

Open-source communication stack that experiments with self-hosted messaging for small teams. The project explores building blocks for channels, permissions, and extensibility so organizations can tailor collaboration tools to their workflows.