Steven M. Ledbetter

(845) 379-0641

NYC Area, NY

Summary

Product leader with 20 years of experience, three acquisitions (Supernatural to Meta, Elevate Security to Mimecast, Habitry to Lift The Bar), and a co-patent on an ML model. I build applied AI with Claude Code daily. I ship open-source agentic development tooling, and I publish research on LLM evaluation and multi-agent systems. My background spans data infrastructure, security, and behavioral science, which shapes how I scope AI products around what technical users will actually adopt.

Employment

Meta

2023 – 2026

Product Manager, AI & Data Infrastructure

  • Co-led the AI & Data Infrastructure org's transition to AI-native, agentic product development, increasing Developer & Data Science productivity by >30% in 90 days.
  • Scaled Meta's real-time metrics and analytics warehouse to absorb 100x agentic-workload traffic in a year (9x in the last 90 days) while tightening latency and freshness SLAs.
  • Drove alignment with internal and external customers and shipped a novel OLAP system with real-time freshness, <100ms latency, 10M QPS, consistency SLOs, 10x capacity savings, and $1Bil+ in new revenue in 6mo.

Supernatural (acquired by Meta)

2022 – 2023

VP of Product

  • Grew the business 3x ARR with 2x user retention with a suite of behavioral science-based features.
  • Guided product through a successful acquisition by Meta and data compliance audits by the FTC.
  • Led a team of 17 PMs & UXR to make what Business Insider called 'the one life-giving product from Meta'.

Elevate Security (acquired by Mimecast)

2020 – 2022

VP of Product

  • First product hire; built the Product Team from scratch including PMs, POs, Designers, and Technical Writers.
  • Pivoted the company to new products that 2x ARR with less than 5% churn before acquisition by Mimecast.
  • Co-inventor on patent US12149553B1, ML behavioral risk-scoring and intervention system to prevent social-engineering attacks.

Tonal

2018 – 2020

Director of Product Management, Engagement

  • Increased engagement 2x and grew NPS to 80+ by implementing Motivation Science education and metrics across the coaching, product, and design teams.
  • Developed the workout experience prior to launch, coordinating across hardware, firmware, software, and content teams.

Habitry (acquired by Lift The Bar)

2013 – 2018

CEO, Head of Product, Co-Founder

  • Co-founded, raised a $150,000 seed, and built the first behavior-change iOS app for health coaches before selling the company.
  • Improved client product engagement 2x and retention 10x using behavioral psychology.
Open Source Projects github.com/smledbetter

Flowstate

https://github.com/smledbetter/flowstate

Observability platform for AI agent development. Tracks metrics, traces, gate-failure signals, drift detection.

  • Sustained 97% prompt-cache hit rate across 130 sprints in 16 projects. Only ~3% of per-sprint tokens are new work, enabling a 842M-token single-operator agent-dev pipeline.
  • Pre-registered factorial experiments; killed 2 of 3 proposed workflow changes after measuring a 10–13% noise floor.

Stipple

https://github.com/smledbetter/stipple

Pip-installable Jupyter widget for AI researchers interacting with 100mil scale datasets faster than anything else on Earth.

  • Repurposes the GPU's rasterization hardware as the predicate evaluator, rendering 10M points at 60 FPS and aggregating 600M+ rows in the browser in less than 80ms.
  • Bidirectional lasso loop: select a cluster, get a uint32 numpy array back in Python, then recolor 10M points at 60 FPS in one GPU buffer swap — a kernel-side primitive jupyter-scatter and Nomic Atlas don't offer.
Writing smledbetter.com/writing

Structure Beats Scale (2026) — Structured review outperforms brute-force generation in LLM code synthesis.

Receipt-Gated Pipelines (2026) — Cryptographic verification of tool-call claims in multi-agent LLM systems.

PM in the AI Era (2026) — Blog series on PM in the AI era — sprint methodology with AI agents, coaching agent teams, and building encrypted apps with LLMs.

Skills

AI & LLM Agents — Anthropic SDK, Claude Code, MCP, eval design, prompt caching, tool-call verification

Product Leadership — three acquisitions, patent co-invention, infra org transitions

Python & SQL — Python, SQL, data analysis, evaluation pipelines

Machine Learning — LLM-as-judge, drift detection, factorial design, noise-floor calibration

Data Infrastructure — DuckDB, OLAP, real-time analytics, telemetry pipelines

Behavioral Psychology — motivation science, behavior change, habit formation

Writing — books, blogs, technical writing

Education

Masters, Sport Psychology — John F. Kennedy University (2014)

Bachelors, Philosophy — University of Chicago (2004)