About

A short story about how I work.

I'm an AI/ML engineer and scientist with a decade of experience taking ambiguous business problems through every step of the lifecycle — framing, data, feature design, modeling, evaluation, serving infra, production rollout, and post-launch iteration — and standing up the platforms that make that lifecycle repeatable.

I started in microwave and electronic engineering, designed wireless and optical telecom systems for oil-and-gas sites and Iran's national mobile network, then turned to machine learning during my Ph.D. at the University of Texas at San Antonio (UTSA). I worked on real-time adaptive perception, anomaly scoring, IoT, and multi-cloud scheduling — research that became a $100K Cisco grant and a U.S. patent on workload-aware job scheduling.

From there I went to Expedia, where I spent six years progressing from MLS to Staff MLS across computer vision, ranking, and recommendation — shipping Expedia's first personalized flight recommender on iOS, the Web Flights LambdaRank LTR ranker, a predictive cache-fill DNN, CTR gating, price forecasting, the contextual-bandit hero-image selector, and a neural ranker and computer-vision stack for Vacation Rentals. After Expedia, I led a multimodal CLIP-based discovery system at Quotograph, then ran the AI/ML org as Sr. Manager / Lead at C5i for Dell's semantic supply-chain retrieval and entity-resolution platform. Today I'm Lead ML Engineer at Procore, working on RAG-and-contextual-intelligence recommendations and an end-to-end LLM Product Insights Platform.

How I lead

// scope · standards · cadence
Cross-functional leadership

I partner directly with executive, product, business, and engineering leaders — defining AI/ML strategy, scoping deliverables, setting roadmaps, and translating ambiguous business needs into concrete technical solutions.

End-to-end ownership

I own model and platform decisions from problem framing through data and feature design, modeling, evaluation, serving infra, rollout, and post-launch iteration.

Standardized ML lifecycle

At C5i I built a repeatable AI/ML guideline — scoping → data → EDA → modeling → eval → deployment → A/B → monitoring — that compressed delivery cadence from annual to quarterly.

People & hiring

I've hired, mentored, and led teams of AI/ML engineers and scientists at C5i, Quotograph, Expedia, and Procore. I design hiring rubrics, interview loops, leveling guidance, and onboarding.

Zero-to-one comfort

I do well in high-ambiguity settings — strong on distributed systems, data infrastructure, real-time / streaming pipelines, and bridging model development with online serving.

External representation

I write up technical work on retrieval-ranking, LLM-based entity resolution, and LLM product platforms, and present at internal leadership reviews and across partner orgs.

Education

// formal training
Ph.D., Electrical & Computer EngineeringAI & Big Data
Aug 2014 – Dec 2017
The University of Texas at San Antonio (UTSA) · Supervisors: Paul Rad, Mo Jamshidi

Dissertation: Real-Time Adaptive Data-Driven Perception for Anomaly Priority Scoring at Scale.

M.Sc., Electrical EngineeringMicrowave & Optical Communication
2008 – 2010
Sharif University of Technology, Tehran
B.Sc., Electrical EngineeringElectronics
2002 – 2007
Semnan University

Honors & awards

  • NSF Graduate Research Fellowship — Grant No. 1419165 (supported Ph.D. research).
  • Partial Ph.D. support by Air Force Research Laboratory & OSD (Grant FA8750-15-2-0116).
  • $100K Cisco grant — IoT + geosocial anomaly detection in smart communities.
  • Open Cloud Institute Outstanding Student Scholarship (2015 – 2017).
  • Lutcher Brown Scholarship and Distinguished Reward, UTSA.
  • Department of Electrical & Computer Engineering Teaching Assistantship, UTSA.

Let's talk

If you're working on transformer-based recommendation, RAG ranking, generative retrieval, or LLM-powered product platforms — and you'd like a senior partner who can own scope end to end — I'd love to hear about it.