Chad Mills
Applying AI & ML to large-scale consumer products
NLP & LLMs · Knowledge Graphs · Ranking · Recommendations · Safety
About
I lead teams building large-scale AI/ML-powered systems. I'm a generalist whose IC time predates today's specialized data roles; I often add the most value in end-to-end alignment — coordinating work across engineering, science, and data, finding the highest-impact problems, and building systems where AI/ML solves the right portion of the problem.
I've been building or leading AI teams for nearly twenty years, across Microsoft, Facebook, Twitter, Grammarly, Primer AI, and Amazon/Twitch. My work spans LLMs, agentic AI, NLP, knowledge graphs, ranking, recommendations, and trust & safety.
I hold a PhD in Computational Linguistics from the University of Washington and an Executive MBA from Quantic. I'm currently based in Austin, TX.
Expertise
AI & ML
- LLMs & Agentic AI
- Natural Language Processing
- Knowledge Graphs
- Recommendations & Ranking
- Deep Learning
- Generative AI
Leadership
- Organizational Design
- Team Building (5 → 35+)
- Cross-functional Alignment
- Goal-setting & Accountability
- Process Management
- Executive Communication
Engineering
- Distributed Systems
- ML Infrastructure
- Data Science
- Analytics
- Algorithms
- Software Engineering
Experience Highlights
Directing 35 people across applied science, engineering, ML infra, data science, and TPM. Grew user base 9% organically in 2 years. Improved ML shipping velocity 70%. Applied multi-objective optimization for 5% revenue gain. 17× increase in annual measured impact.
Led 4 teams building Twitter's knowledge graph. Systems directly responsible for 10% of Twitter's revenue, contributing to over two-thirds of revenue through dozens of cross-company partnerships. Led director/VP-level initiative enabling personalization across all three divisions.
Organized 35 ML engineers, researchers, and linguists into 5 outcome teams. Scaled team from 5 to 35. Shipped dozens of ML, LLM, and Transformer-based GenAI features. Increased upgrades 10% with tone detection — the most impactful feature in Grammarly's history.
Technical leadership for core product at Series B startup, used by CIA analysts. Grew revenue from $1M/yr to $10M/yr. 1200× performance improvement through system rearchitecture. Reduced query response from 20 minutes to sub-second.
Increased user engagement with News Feed by 5%. Established standard experiment review process. Drove executive reporting identifying failing segments and strategic opportunities, leading to 3 major new initiatives.
12 years across trust & safety, anti-spam, and anti-malware. Decreased socially-engineered malware two orders of magnitude. Led industry in browser protection — NSS Labs found 96% less malware missed than Google. Decreased spam from 50%+ of inboxes to under 4%.