Looking for a job

Gabriel Payan

@gabrielpayan

Senior AI/ML Engineer at Bloomreach who cut p95 reranking latency from 140 ms to 85 ms.

At Bloomreach, I worked on ecommerce product discovery, combining Elasticsearch retrieval with learning-to-rank and embedding-based candidate generation. I reduced p95 reranking latency from approximately 140 ms to 85 ms in the evaluation workload by caching stable features and batching inference calls.

I also developed retrieval-augmented generation workflows for conversational search and built offline evaluation datasets to compare ranking candidates. At Constructor, I developed search and recommendation features, while earlier work at Salsify and Narvar included catalog data pipelines, validation, and analytics preparation.

Experience

Work history, roles, and key accomplishments

Senior Machine Learning Engineer

Bloomreach

Jun 2023 - Aug 2026 (3 years 2 months)

Developed product discovery services with Elasticsearch-based retrieval and learning-to-rank models, improving ecommerce search relevance. Built offline evaluation datasets and reduced p95 reranking latency from 140 ms to 85 ms.

Elasticsearch Learning To Rank LightGBM PyTorch OpenCV RAG NDCG@K MRR Interleaving Python AWS

Machine Learning Engineer II

Constructor

Jan 2021 - May 2023 (2 years 4 months)

Developed ranking features and recommendation models for ecommerce search, addressing sparse behavioral data and cold-start scenarios. Reduced p95 recommendation response time by 22% through profiling and caching.

Python NodeJS Recommendation Systems Collaborative Filtering Feature Engineering Caching Profiling Elasticsearch

Machine Learning Engineer

Salsify

Oct 2018 - Dec 2020 (2 years 2 months)

Developed Python and SQL pipelines to normalize product data for catalog search, reducing indexing failures. Experimented with text similarity and classification models for category suggestion and duplicate detection.

Python SQL Elasticsearch scikit learn Data Pipelines Semantic Similarity Classification

Software Engineer, Data & Machine Learning

Narvar

Jan 2018 - Sep 2018 (8 months)

Maintained Python and SQL jobs for ecommerce shipment data, adding deduplication and timestamp checks. Supported feature preparation and evaluation for shipment-status prediction experiments.

Python SQL scikit learn Data Pipelines Deduplication Logging