* Off-policy Evaluation* Non-uniform Random Exploration and Supervised Learning

* Retriever-Generator (RAG) QA Systems* LLM Generative Pre-training* LLM Tuning* ANN Search & Tuning

* Fine-tuned BERT on domain-specific Ranking Task* Billion-scale Cost-efficient BERT Ranking via Amortized Inference: +4% Revenue* E2E Deep Retrieval (candidate-free ranking)

Personalization and Search Research exploring Representation Learning, Factorization, Sequential and Deep Models for Recommendations.* Restaurant Recommender (Factorization-based with side-information)* Menu Item Recommender (Deep Next-Item model w/ contextual-info)* Personalized Search using LTR techniques (Pairwise data, RankNet, Query Expansion)* Contextual Bandits for Marketplace Diversity* Semantic Search using 2-stage IR system* Search Intent Classification… Show more Personalization and Search Research exploring Representation Learning, Factorization, Sequential and Deep Models for Recommendations.* Restaurant Recommender (Factorization-based with side-information)* Menu Item Recommender (Deep Next-Item model w/ contextual-info)* Personalized Search using LTR techniques (Pairwise data, RankNet, Query Expansion)* Contextual Bandits for Marketplace Diversity* Semantic Search using 2-stage IR system* Search Intent Classification using Weak Supervision* Offline AB Testing via Counterfactual EvaluationTalks & Publications* Accepted @ RecSys 2021: Online Learning for Recommendations at Grubhub* Accepted talk at GTC 2021: Counterfactual Learning to Rank in E-commerce* Invited Talk at Tensorflow Ranking Workshop 2020* Accepted Talk at PyData 2019: Latent Food Graph* Feature Blog: Query Expansion using query2vec @ GH* Featured Blog: Weak Supervision @ GH Show less

Engaged in research and development of computer vision solutions for Hotel Search Products* Hotel ranking algorithm based on image data. 5% CTR increase (CNNs, Siamese Networks)* Image Based Recommender System. 2% CTR increase (Bayesian Personalized Ranking, CNN)

Lead data initiatives for hedge fund. Lead design, development and deployment Default Forecasting and SEC Filings Search products* Loan default forecasting product (Tree-based)* SEC filing search system (Distributed Computing, Elasticsearch)* Cloud migration: 70% cost savings

Lead technical efforts for food logistics company. Product launches in San Diego and Denver.* Technology improvements to consistently grow sales by +25% month-to-month from 10/2014.* Migrated to load adaptive servers cut hosting fees by 75%.* Automated daily email system (90% throughput increase).* Automated SMS ordering system (3% conversion increase)* Automation decreased avg. delivery time by ~30 min (more sales/less labor)* Recommender System (5% conversion… Show more Lead technical efforts for food logistics company. Product launches in San Diego and Denver.* Technology improvements to consistently grow sales by +25% month-to-month from 10/2014.* Migrated to load adaptive servers cut hosting fees by 75%.* Automated daily email system (90% throughput increase).* Automated SMS ordering system (3% conversion increase)* Automation decreased avg. delivery time by ~30 min (more sales/less labor)* Recommender System (5% conversion increase, Collaborative Filtering on Implicit Feedback) Show less

Lead engineering effort to build on-demand real-estate showing platform (“Uber for Real Estate”). Product Development and Engineering. Successfully raised 1M+ seed funding to build product from prototype, to MVP, to production rollout with over 2000 live paying customers.* Developed Real-time evented chat-systemAcquisition in July 2015

Lead initial product prototype and development. Helped grow MVP to product and 1M+ seed funding and acceptance into evonexus accelerator. Lead small team to scale out platform to $3000+/month sales.* SaaS Churn Modeling* 2014 - Current: Advisory Role

Tasked with commercializing R&D IP into products* Gimbal - R&D work with proximity beacons (ios/Android) - QC Exit and Acquisition* Connected Home Project - Embedded Programming / Hardware Integration* Augmented Reality & Computer Vision Projects (QCAR/Vuforia)* Tapioca Mobile - Application Level development (Ruby, Rails, TDD, Git)

Created platform for mobile video distribution. Was acquired by Qualcomm.

Hardware Integration / Software Development: USB interface w/ in-house hardware platform. Real-time 3D modeling w/ DirectX.

QIS/Brew/Tools & Automation TeamDeveloped software w/ small team. Planned and Lead projects related to the Brew Platform.

Developed reporting and analytics software. Developed automated testing application for BREW apps. Was acquired by Intellisync corp, then Nokia.