Lead Data Scientist // Production AI
Selected systems, public code, and recent shipping activity.
Portfolio index // Shipped systems // Public work
Lead Data Scientist building production agentic AI systems with a focus on practical, reliable products
A mix of production systems and public builds, selected for impact, technical depth, and inspectable results.
Enterprise-scale LangGraph and LangSmith platform that turns PDFs, emails, and spreadsheets into pricing bids across $80B+ in annual freight workflows.
Outcome: Cut turnaround time by 40 hours and freed 20 FTEs.
High-performance time-series forecasting library written in Rust. A Prophet reimplementation running in production for mission-critical decision systems.
Outcome: Predictions run 15–20× faster than the Python implementation.
Interactive TypeScript project for comparing lump-sum investing and dollar-cost averaging across real SPY history.
Focus: Break-even distributions, recovery stress tests, and clear decision framing.
A public-only snapshot of meaningful work across recently active projects. Updated July 26, 2026.
8806622
286455e
def73c0
5b130aa
d0acc97
3a43a15
68a29ee
5d999a3
Public repositories only // Authored commits // Static snapshot
I build AI systems and production products that solve real problems. Over the past 6+ years, I've worked across the full spectrum—from designing intelligent agents that automate complex workflows, to developing high-performance time-series infrastructure, to creating tools that help people make better decisions with data.
My work sits at the intersection of research and engineering. I take emerging techniques in AI and machine learning and turn them into reliable, scalable systems that deliver value. The common thread is building production-grade software that people actually use and trust.
Design and deploy multi-agent platforms using LangGraph, LangSmith, and RAG. Built enterprise systems handling $80B+ in annual business processes with 20+ FTEs freed.
Deploy ML systems on Kubernetes that actually scale. Rewrote critical forecasting libraries in Rust for 10x performance improvements in production pipelines.
Build conversational analytics agents that let business users query complex data warehouses in natural language and get context-aware insights.
Set up CI/CD pipelines, Kubernetes deployments, and observability frameworks for production AI systems on GCP and AWS.
Build costing engines, pricing algorithms, and predictive models that form the backbone of business strategy and drive real ROI.
Create "trust but verify" frameworks for production AI, handling model validation, observability, and responsible deployment of business-critical systems.
Exploring the challenges and interesting discoveries from reimplementing Facebook's Prophet forecasting library in Rust.
Read More →Interested in agentic AI, production ML systems, or building AI that drives real business value? Let's connect on LinkedIn or check out my code on GitHub.