Live web product · Equity research
An automated equity-research platform that turns company data into structured, analyst-style reports. The product combines financial data pipelines, generated commentary and a focused editorial interface designed for rapid company analysis.
View live projectPrivate system · Quantitative execution
A quantitative trading tool built for Interactive Brokers, pairing Python research flexibility with Cython-accelerated components. Designed around market data, strategy logic, risk controls and broker execution in one cohesive workflow.
Private systemMachine learning · Language model
A 303-million-parameter language model created in Python with PyTorch. The project covers the model-building workflow as an engineered system: data handling, training infrastructure, model execution and iterative experimentation.
Research projectThe common thread is end-to-end ownership: turning an idea into data pipelines, models, interfaces and working systems rather than stopping at a notebook or prototype.
Data ingestion, analysis, model logic, execution and presentation are designed as one connected system.
Use expressive Python for iteration, then move performance-sensitive components into more efficient implementations when justified.
Technical depth should not produce confusing products. Interfaces are structured to make models, signals and research understandable.
Ideas are treated as hypotheses to test, measure and refine.
Open to conversations around quantitative research, financial technology, machine learning and ambitious systems work.