Origin browser workspace screenshot

Origin

Co-founder · LLM-powered browser workspaces

Feb 2023 – May 2023
  • Co-founded a full-stack platform using LLMs and NLP to curate context-aware personalized browser workspaces; received a seven-figure valuation from Pear VC.
  • Worked with semantic search, recommendation algorithms, fine-tuned chatbots, and clustering.
Ephemeral logo

Ephemeral

Autonomous voice-driven AI agent

Feb 2024 – Mar 2024
  • Built a real-time AI assistant that listens during meetings and proactively drafts emails, schedules calendar events, retrieves answers, generates images, and more.
  • Combined speech-to-text/text-to-speech, audio processing, supervised classification, RAG, few-shot learning, multi-threading, and data augmentation.
BiteBuddy logo and tagline

BiteBuddy

All-in-one meal planner & personal CRM

Oct 2023 – Nov 2023
  • Created a full-stack platform using ML to figure out what to make for every meal and to maintain personal relationships.
  • Used spectral clustering, RAG, recommendation systems, and graph ML.
Online Interlingua-English translator screenshot

Online Interlingua Translator

First public, open-source Interlingua–English translator

May 2020 – Jan 2023
  • The first free, online, public, open-source Interlingua and English translator.
  • Provides a UI similar to other online translators like Google Translate.
  • Used by 5,000+ people.
Interlingua translator Colab notebook screenshot

Interlingua Translator (Colab)

Source notebook for the Interlingua-English model

May 2020 – Jan 2023
  • Open-source training and inference code for the Interlingua–English translation model.
  • BLEU score of 42.45 (EN→IA) and 42.54 (IA→EN), "high-quality" by Google's AutoML guidance.
Blackjack game board screenshot

Blackjack in Java

Standard Blackjack implementation

Nov 2021 – Dec 2021
  • Designed a command-line implementation and GUI of standard Blackjack in Java with a focus on design patterns.
  • Implemented model–view–controller and observer patterns.
Interlingua Corpus Project logo

Interlingua Corpus Project

Largest corpus of Interlingua sentences

May 2020 – Jan 2023
  • 1.2M+ quality-controlled Interlingua sentences scraped via a custom web crawler.
  • Visited 6M+ relevant websites.
  • Provides 80,000+ parallel Interlingua-English sentences and other useful data.