Lionel Eisenberg

San Francisco, CA, 94107  •  [email protected]  •  github.com/LionelEisenberg  •  LinkedIn  •  Green Card Holder

Engineer who treats infrastructure as a product: built data pipelines and ranking systems behind Google Search and a 1M-page publishing platform at Yext, owning outcomes — not just uptime. Outside work: a self-hosted home & fitness tracking platform, LLM tooling, a game in progress — mostly excuses to learn something new.

Professional Experience

Google San Francisco, CA
Software Engineer — Search Team / Merchant Content & Discovery Team 06/2022 – 05/2025
  • Developed a Merchant Targeting Engine API to provide Google surfaces (Search, Gmail, Ads) with real-time access to comprehensive merchant information (type, attributes, categories). Leveraged high-availability storage and MapReduce pipelines to process data from across Google into structured summaries served to search and consumer surfaces. Led to an increased successful onboarding rate for merchants of 22% over 2 years.
  • Built a service for direct merchant data synchronization between Google Shopping and 3rd-party ecommerce platforms (Shopify, WooCommerce). Developed a modular scraping agent to identify and categorize ecommerce stores, and an API initiating secure 3P and Google OAuth flows for seamless authentication and data sync. Increased successful onboarding rate of 3P Merchants from 12% to 44%.
  • Revamped the “hidden gem” ranking component within Google Search, integrating new signals to surface high-quality, lesser-known merchants for shopping queries; the quality module was redesigned and retrained in collaboration with Search stakeholders. Contributed to 46% of a top-level Google OKR by diversifying search results.
  • Spearheaded “New Arrivals”, a product detection system, tagging products sold on Google as New Arrival, Trending, or Promotion with no merchant action. Daily rolling MapReduce pipelines and fuzzy matching over SKUs, categories, and attributes decide when two listings are the same product. Led end to end, from spotting the unused product data to launch; increased new-product discovery by 18%.
  • Served as a Java Readability Mentor: conducted code reviews, interviewed prospective candidates, ran monthly team retrospectives, and mentored interns and junior engineers.
Yext Washington DC
Software Engineer 06/2020 – 06/2022
  • Built a high-throughput static asset generation and publishing system capable of publishing 1M pages in under 20 minutes, each accessible in <30ms from CDNs globally.
  • Led the creation of a CLI tool harnessing the page generation logic, empowering developers to build, serve, and render pages locally given mock data sources.
  • Architected a Nomad GroupeCache between Redis and Nomad clusters — increased cache hit rate by 135% and lowered average request time by >40%.
  • Served on the monthly on-call rotation ensuring 99.99% uptime, responding to client requests and PagerDuty; designed and implemented technical interviews and system-design challenges for prospective applicants.
  • Tech Stack: GoLang, AWS S3, GCS, Kafka, MirrorMaker, Kubernetes, MySQL, Redis, Route53, Nomad
ForagerOne Baltimore, MD
Software Engineer www.foragerone.com 09/2017 – 05/2019
  • Managed a team of 10 engineers building a full-stack platform that helps students discover faculty-affiliated research opportunities. Shipped messaging, profile UIs, and backend features, reduced technical debt, resolved critical failures in real time, and mentored junior engineers in incident response.
  • Tech Stack: Angular, NodeJS, MySQL, Firebase, Elasticsearch, Express

Projects

lionel.place lionel.placegithub.com/LionelEisenberg/Pirateship Personal Project
  • Built “Helm”, an AI-driven health & fitness tracking app, to product standard (FastAPI, React/TypeScript, ~33k LOC, 600+ tests gated by pre-commit). Logs meals, workouts, sleep, and habits from free-form text or food photos via LLM parsing, tracks progressive overload across workouts, and syncs Google Health data including sleep stages and intraday heart rate. Built as a consumer product; used daily by me, friends, and family.
  • Built an asynchronous LLM pipeline: a durable job queue consumed by a dedicated worker container (leases, retries, idempotency keys), with per-task model waterfalls across Claude (Agent SDK) and Gemini, quota tracking, and automatic failover to self-hosted Ollama models on a local GPU when both cloud providers fail. Every job records token, cost, latency, and confidence telemetry, and clients reconnect to in-flight jobs after a page refresh.
  • Self-host and operate a 17-service Docker stack: media streaming and requests, GPU-accelerated photo library with ML search, password manager, and a blog with email subscriptions, fronted by Caddy (automatic HTTPS, HTTP/3) with Cloudflare dynamic DNS and Jellyfin-backed JWT auth with role-based access.
DenyTillYouDie github.com/LionelEisenberg/DenyTillYouDie Kearny Game Jam 2025
  • Satirical dark-humor healthcare simulator: you play a claims assessor forced to choose between approving life-saving care and hitting the quotas that keep you paid, employed, and able to feed your family.
  • Designed, built, and shipped solo within the jam window: all gameplay systems, UI, and balancing in GDScript; all pixel art drawn in Aseprite.
EndlessPath github.com/LionelEisenberg/EndlessPath Big Mode Game Jam 2026
  • Incremental idle RPG with active mini-games set in a world of Wuxia myth — you slowly cultivate a martial artist's power.
  • In development since late 2025, the bigger of the two games: idle and progression systems, mini-games, custom GDShader effects, and all art in Aseprite.

Education and Diplomas

Johns Hopkins University Baltimore, MD
Graduated with honors, Cumulative GPA: 3.85
Master of Science and Engineering (M.S.E) in Computer Science, focus in Cryptography 06/2020
Bachelor of Science (B.S) in Computer Science, Minor in Applied Mathematics & Statistics 06/2019

Skills

Programming Languages:
Java, GoLang, TypeScript, Python, SQL, GDScript, GDShader, Dart, C++
Backend & Infrastructure:
FastAPI, SQLAlchemy, AWS (S3, Route53, SQS, Lambda), Kafka, Nomad, Kubernetes, Docker, Redis, MySQL
LLM Systems:
Claude Agent SDK, Anthropic & Gemini APIs, multi-provider routing & failover, structured outputs, async job queues, LLM observability (token/cost/latency telemetry)
Machine Learning & Data Science:
TensorFlow, Ranklab, Anima, Objective-Based Tuning
Game Development:
Godot (GDScript), C++, OpenGL, shader programming (GDShader), Aseprite pixel art, 2D game & systems design
Languages:
Native French & English; operational Mandarin & Spanish