Principal ML Engineer Agentic AI · Ranking · Decision systems
AI systems.
Built for reality.
I turn advanced machine learning into dependable products—connecting model capability to architecture, evaluation, and measurable consequence.
15+ years across agentic AI, consumer-scale ranking, reinforcement learning, and distributed systems.
Selected evidence
Impact measured in deployed systems, not prototype claims.
01 / Engineering thesis
The model is not the product.
A reliable AI product is a decision system: models, data, orchestration, policy, evaluation, interfaces, and feedback working under real constraints.
Capability
Can the model reason, predict, or act well enough?
Reliability
Can the system behave predictably outside the happy path?
Consequence
Does the product improve a decision that matters?
02 / Selected systems
A career viewed through engineering problems—not job descriptions.
Three systems. Three kinds of consequence.
Turning agent capability into a dependable product system.
The engineering challenge is larger than model intelligence: orchestration, routing, evaluation, policy, observability, and recovery have to behave as one system.
Engineering consequenceFrom impressive paths to reliable operating envelopes
Representative decision loop04 stages
01Intentcontext + goal
02Orchestrateroute + plan
03Actmodels + tools
04Evaluatequality + safety
↳ Feedback
Production signals feed evaluation, routing, and recovery design.
03 / Operating model
How I move from ambiguity to an operating system.
Principal-level work is not only choosing an architecture. It is creating the conditions for a system—and a team—to make better decisions repeatedly.
Frame the decision
Begin with the decision, the user, and the operating constraint—not a preferred model.
Design the system
Treat data, models, orchestration, latency, policy, and human trust as one engineering surface.
Prove the behavior
Use evaluation and observability to understand where the system works, fails, and recovers.
Scale the learning loop
Build feedback into both the product and the organization operating it.
2026—Now
PayPal
Principal ML Engineer · Agentic AIProduction-grade agentic experiences and the systems discipline behind them.2024—2026
Meta · Instagram
Machine Learning EngineerRelevance and ranking for consumer-scale product experiences.2019—2024
Walmart Global Tech
Principal Engineer · Senior Manager II · Staff ML EngineerApplied ML spanning deep reinforcement learning, pricing, and large-scale decision systems.2015—2019
Intuit
Senior Software EngineerLarge-scale data and software systems for high-trust financial products.2009—2014
HSBC
Senior Software EngineerDistributed systems foundations for complex, regulated environments.05 / Research & recognition
Ideas tested where the stakes are real.
INFORMS · Published research
A Multiobjective Optimization for Clearance in Walmart Brick-and-Mortar Stores
Deep reinforcement learning, simulation, and optimization shaped into a deployed decision system.
Read the paper ↗Recognition2020Franz Edelman Award laureate
Outstanding achievement in advanced analytics.
INFORMS ↗Research profileG
Google Scholar
Publications and citations across applied machine learning.
View profile ↗06 / Connect
Building machine intelligence that earns trust in the real world.
For ambitious product and platform problems across AI agents, machine learning, and large-scale systems.