Hector Romero

S

Senior Data Scientist at Nvidia · Recommendation systems · PyTorch, Spark

Nvidia
Seattle, WA · PT (UTC-8)
5+ years experience
University of Washington
English
Co-founder
Messaging unavailable

Overview

I work on recommendation systems at Nvidia in Seattle. Moved the homepage from rules to a learned ranker. Shipped a ranking change that increased conversion 4.6%. Before that: Verizon and Dropbox. Collector of fountain pens.

Good reasons to reach out
  • People who want to build this with me
  • Thoughtful intros to people in my field
  • Questions about recommendation systems
Please don't message about
  • Offshore staffing offers
  • Agency or outsourcing pitches
  • Survey or panel invitations

Travels often; replies within 48 hours

Skills
PythonSparkPyTorchrankingembeddings
Co-founderLooking for a co-founder
closes in 3 months

Need an operations co-founder: cold-chain monitoring in the UK

The idea
Need an operations co-founder: cold-chain monitoring in the UK
Stage
Customer interviews
I bring
14 years in recommendation systems
Looking for
Operator who has scaled a team
Commitment
Full-time now
Location
Seattle or remote

I bring 5 years of recommendation systems (currently Nvidia). Ideally you've run operations in India. Have a working prototype and design partners.

Seattle, WApytorchsparkpython
Open

Experience

  1. Senior Data Scientist
    Feb 2026 - Present
    Nvidia · 8 mo · Full-time · Seattle, WA
    • Shipped a ranking change that increased conversion 6.9%.
    • Wrote the metric definitions the leadership team reviews every week.
    • Taught SQL to 194 non-analysts.
  2. Data Analyst
    Jan 2024 - Feb 2026
    Verizon · 2 yr 1 mo · Full-time · Seattle, WA
    • Moved the homepage from rules to a learned ranker.
  3. Data Scientist
    Jun 2022 - Jan 2024
    Dropbox · 1 yr 7 mo · Full-time · Seattle, WA
    • Shipped a ranking change that increased conversion 8.5%.
  4. Data Scientist
    Oct 2021 - Jun 2022
    Walmart · 8 mo · Full-time · Seattle, WA
    • Moved the homepage from rules to a learned ranker.
    • Taught SQL to 82 non-analysts.

Projects

Conference talk: lessons from recommendations
2025

What went wrong and what worked while scaling recommendation systems.

Side project: a reading list app
2022

Built with Spark and PyTorch. 1682 people used it last year.

References

No references yet.

Education

  1. University of Washington
    BS, Physics
    2017 - 2021

Details

Compensation expectations
What they are targeting for their next role.
Visible to connections
Direct contact details
Direct email and a scheduling link.
Visible to connections

Trusted Partners

Sponsored