← Kevin Lee

Resume

kevincorvallis@gmail.com github.com/kevincorvallis

I build at the intersection of scale and intelligence. Today I lead the Applied ML team for Short Form at Paramount, architecting production AI pipelines — multimodal embeddings, semantic retrieval, classification — across one of the world's largest media catalogs. Before that I shipped Microsoft Copilot to millions, mapped brain regions at the Broad Institute, optimized the Curiosity rover at NASA JPL, and served five years in the U.S. Army. On weekends I fly medical transport for Angel Flight West.

Experience

Paramount — Software Engineering Manager, Applied AI/ML

Lead the Applied ML team for Short Form, building end-to-end AI workflows for content classification, multimodal embedding, semantic retrieval, and personalization. Architected a Google Cloud pipeline that turns raw media into searchable intelligence — Gemini multimodal embeddings, LLM taxonomy classification, output validation — powering a clip-classification platform (FastAPI · Cloud SQL · Pub/Sub · GKE) at ~2s p50 and ~$0.0003 per clip, replacing manual editorial review at catalog scale.

Microsoft — Software Engineer

Shipped the new Microsoft Copilot settings experience end-to-end on Azure, serving millions of daily users. Designed and operated Cosmos data pipelines ingesting 40TB+/day across Copilot, Edge, and Windows (100M+ monthly users), cutting experimentation latency from 72 to 24 hours; migrating always-on VMs to event-driven Azure Functions trimmed $240K+/yr in compute at 99.9%+ uptime. Led a 12-person team to win the 2022 Microsoft Global Hackathon with an AI tool for early Alzheimer's detection.

Broad Institute of MIT and Harvard — Researcher

In Dr. Guoping Feng's lab, classified spatial scRNA-seq (MERFISH) to subthalamic nuclei regions using machine learning on DAPI/poly-T scans from the Allen Brain Institute.

U.S. Army — Soldier

Five years of service. Led and trained squads of 8–12 through tactical field exercises, then ran logistics as Battalion S-4 for a 500+ person battalion. Full Army ROTC scholarship.

NASA Jet Propulsion Laboratory — Software Engineering Intern

Built scheduling-optimization models for Curiosity rover operations, improving mission-planning efficiency 8–12% and cutting idle time between science activities. Presented the work at the 2020 NASA conference.

Selected Work

Side projects shipped under my studio, PookieB — across web, iOS, and AI/ML.

Shredders — real-time powder tracker, 53 resorts, 90 API routes, built on Next.js.

SPACEc — Stanford / Nolan-Lab Python workflow for multiplexed-imaging analysis, published in Nature Communications; on PyPI.

SmartSpender — AI financial-decision engine, a 3-stage pipeline across 13 AWS Lambdas.

See all projects →

Education

University of Southern California — B.S. Computer Science, Viterbi School of Engineering · 2022

Honors — Army ROTC & Pathway Oregon full scholarships · Armed Forces Award · Dean's List

Volunteer

Angel Flight West — Pilot

Volunteer pilot flying medical transport across the Puget Sound for Angel Flight West.

Skills

Languages: Python, C++, C#, TypeScript, Java, SCOPE

AI / ML: Gemini, multimodal embeddings, LLM classification, PyTorch, TensorFlow

Cloud & Infra: GCP (Vertex AI, GKE, Cloud SQL, Pub/Sub), Azure (Functions, Cosmos), AWS (Lambda, DynamoDB), Kubernetes, Docker

Frameworks: FastAPI, Next.js, React, SwiftUI, Node.js

Interests

Humanitarian aid · sustainable energy · surfing · flying · traveling


Ask me anything

Want to ask about any of this? Ask my AI — it answers as me.


Let's connect

Want to connect? Book a time that works for you.

Book an appointment