$ ls -al experience/work/

Engineering on Amazon MSK, AWS's fully-managed Apache Kafka service — distributed streaming infrastructure run at scale.

Shipped customer-facing automated limit-increase approvals and a daily Lambda that purges non-compliant log groups across every account in the service.

Built an end-to-end ML pipeline forecasting customer LTV and CAC, with Snowflake tasks pushing results into a Looker dashboard for stakeholders.

Core contributor to Manify, a non-Euclidean representation learning library (75+ stars). Implemented every curvature-estimation method with full tests and docs; now working on a follow-up characterizing the distribution of delta-hyperbolicity.

Implemented GPU-accelerated kNN mutual-information estimation over relational joins without materialization, for SUNA — a scalable causal confounder-discovery system.

Explored dimensionality-reduction theory, then pivoted to agentic data analysis and optimized deployment of unsupervised algorithms in a web app.

Adapted Prior-Fitted Networks (a transformer architecture for few-shot Bayesian inference) to EHR data by reworking the synthetic data-generation process.