Selected work across machine learning, deep learning, probabilistic modeling and production engineering, with an increasing focus on model development and experimentation.
2026 · Independent project · Deep learning
Implemented a character-level decoder-only Transformer in PyTorch from first principles to understand the mechanics of modern language models beyond high-level libraries.
The project includes the training data pipeline, causal self-attention, multi-head attention, residual blocks, evaluation, checkpointing and autoregressive generation.
View on GitLab2025 · Independent project · Probabilistic ML
Built a probabilistic modeling pipeline for ATP match outcomes and evaluated its predictions against bookmaker markets.
Rather than focusing only on classification accuracy, I evaluated calibration, Brier score and decision quality, including Kelly-based analysis of when predicted probabilities provided useful signal.
View on GitLabPrivate client · Production project
Designed and built a production-facing analytics application for a private client, combining product catalog and sales data into a single workflow for understanding commercial performance.
Built ingestion and transformation flows, catalog-to-sales matching, validation logic, analytics calculations and interactive product, customer and order-level views.
Delivered the project as a usable production system rather than a notebook prototype, including data refresh workflows, backend logic and handling of real-world data inconsistencies.
Worked end to end from understanding the underlying business questions through implementation and iteration, keeping the product focused on useful decisions rather than generic dashboard metrics.
2025–now · Livesport
Work in a production product environment across machine learning, analytical problems, data pipelines and process automation.
The role has given me experience working with real production systems, large-scale data workflows and the engineering constraints that sit around applied machine learning.
My independent work is increasingly focused on deep learning, model development and research-oriented experimentation. I’m particularly interested in understanding models from first principles, reproducing ideas from papers and building experiments that explain why systems work or fail.
I’m open to selected remote part-time contract work with small ML and AI teams, particularly roles involving model development, experimentation and technically challenging engineering.