Sunnyvale, CA · aamodsathe.com
Aamod Sathe
Data & AI leadership — I build analytical functions from nothing, scale them, then make them AI-native before anyone asks.
20+ years across Google, Meta, PayPal, and Vanta. Player-coach: comfortable in the CEO's business review and writing the SQL when it matters.
eBay · PayPal
Facebook
Google
Meta
V
2005–2013
2013–2015
2015–2020
2020–2026
2026–Present
20+
Years in data & AI leadership
60+
Data scientists led globally at Meta Reality Labs
80%+
Data questions automated at Vanta via MCP
$1B+
Incremental revenue driven at Meta Ads
Experience
VantaMay 2026 – PresentCurrent
Head of Data & Analytics
- Built MCP-based self-service layer; 80%+ of company data questions answered without analyst involvement at near-100% accuracy
- Built LLM transcript parsing pipeline end-to-end — annotation framework, ground truth baselines, fine-tuned model at 88–92% accuracy at a fraction of prior API cost
- Owns CEO's weekly metrics forum; rebuilt company KPI framework and full data architecture on dbt + BigQuery in six months
- 40% headcount growth in four months; highest manager favorability gain at this level company-wide
MCPLLM EvaluationdbtBigQuerySemantic MetricsCEO Partnership
1 / 8
Building
Four AI products built independently using Claude Code — all active, all with real users. The site itself is a demonstration of that.
Samani.ai
AI interview prep platform for candidates and recruiters. Custom evaluation, role-specific criteria.
Active · 100s of users
Kate
AI phone answering service for frontline businesses. Handles calls, books appointments, routes inquiries.
Active · 100s of users
Coach Samani
AI school exam prep tutor for students. Adaptive practice and personalized feedback.
Active · 100s of users
DataMod Ventures
AI-native venture and investment research platform. Investing in India's AI revolution.
Active
Writing
On AI, data leadership, analytics culture, and building things.
Why I build things with Claude Code on weekendsComing soonThe Portal deprecation: what a data-driven product decision actually looks likeComing soonThe executive metrics forum: the most underrated analytical function in a companyComing soonBuilding a semantic metrics layer: why the data contract is the productComing soonFrom 12 to 50 DS in a year: what I learned about scaling analytical teamsComing soon