Pablo Felgueres

Software engineer living in San Francisco. I've worked on ML systems at Uber, interfaces for language models at Perplexity, and performance simulations for some of the largest solar plants in the US.

In 2022 I left my job to learn how to build products fullstack - design, engineering, marketing, and sales. Highlights have been an LLM playground (top 10 on Product Hunt), a growing app for SF residents to simplify parking (featured on ABC news), and building an engineering community around the future of American manufacturing.

In my free time I enjoy reading - lately bios of great adventurers and US history. I'm a regular open water swimmer in the bay.

Email: pablofelgueres@gmail.com@felgueres on Github@pfelgueres on XMost people don't think simple enough.
Jim Keller
pfp

Street Cleaning Parking (2024)

[link]

Simplify parking in San Francisco.

  • Detects parking rules from geolocation
  • Simple useful design
  • Notifications & calendar integrations
ios scpios scp

AI Answers with widgets (2023)

[link]

Combine RAG with widgets to answer queries.

  • LLM-based query classifiers
  • Productized at Perplexity AI
ios scp

Autodoc (2023)

Fast data extraction for document intensive businesses

  • Built for an insurance startup in India (LoopHealth)
  • LLM pipeline
ios scpios scp

LLM playground (2023)

[link]

Tools to run language tasks on user content.

  • Top 10 product of the day on Product Hunt with 700 active users
  • Multi-turn conversations with citations and chat history
  • Headless browsing to crawl sites
  • Speech recognition models to transcribe user videos
ios scpios scp

Product Engineering SDK (2022)

Tools to build better products faster.

  • Dynamic JSON configs to control code remotely
  • Feature flags and A/B testing
  • SDK to capture events and analytics