Dr. Heather Leffew
Implicit Signal Measurement·Human-AI Behavioral Science·Research Leadership

Dr. Heather Leffew

Measuring what machines and humans reveal through language, behavior, and implicit signals to build the science of how they think, interact, and shape each other.

Background

Biography

Dr. Heather Leffew is a PhD psychologist and data science executive who has spent twenty years building measurement systems that cut through ambiguity in human and machine behavior. Her career is built on a single obsession: using language, behavior, and implicit signals to produce real measurement in domains defined by ambiguity. From forensic psychiatric evaluation to billion-record platform safety at TikTok to building Spokeo's AI function from the ground up, every role has been a different implementation of the same mission: understand how humans and machines think, interact, and influence each other, and build the infrastructure to make that understanding rigorous, scalable, and actionable.

Dr. Heather Leffew
Academic Background

Education

Doctor of Philosophy (PhD), Psychology

Fielding Graduate University / APA Accredited

Specialization
Quantitative Predictive Linguistics
Dissertation
Instrumental and Affective Mass Murder: Establishing a Predictive Typology with Computer-Mediated Linguistic Analysis.
Conference Presentation
Implicit Power Drives in the Manifestos Preceding Autogenic Massacres.
Organizational Outcomes

Professional Experience

Director of Data Science

Spokeo

Sep 2025 - Present

Founding technical leader who built and scaled Spokeo's Data Science and AI function, delivering production ML and NLP systems across search relevance, fraud detection, customer intelligence, and data procurement in partnership with ~80 engineers and product leaders across 8+ cross-functional squads.

  • Owned the AI operating model with executive leadership, defining an OKR taxonomy and delivery mechanisms to translate strategy into measurable execution.
  • Led end-to-end delivery of a proprietary 15-step ML intelligence pipeline combining knowledge graphs, graph ML, and weak supervision for entity resolution and data fusion across ~49TB of multi-vendor data.
  • Set production standards for scale and iteration in partnership with engineering, reducing core processing time from 150+ hours to 24-36 hours and establishing A/B testing and model evaluation practices to support reliable, repeatable releases.
  • Improved data procurement decisions by implementing source independence and ablation frameworks to quantify vendor redundancy and ROI.
  • Delivered NLP-driven schema discovery via a semantic clustering engine that mapped thousands of vendor columns to 75 canonical attribute types, raising schema discovery coverage to 72%.
  • Embedded Responsible AI and a 4-tier regulatory compliance system into the ML lifecycle for compliance-ready deployment.
  • Developed recruitment strategy and nurtured a high-performing data science team, reinforcing technical rigor and strong cross-functional partnership.

Full-Stack Principal Data Scientist & Applied Researcher

TikTok USDS | ByteDance

Jul 2022 - Jul 2025

Owned the roadmap for ML measurement and detection capabilities and led cross-functional teams up to 25 (data science, engineering, policy), setting evaluation standards for accuracy, safety, and operational readiness.

  • Launched an analytics product suite (text analysis pipeline + multi-signal detection frameworks) processing 1B+ records, improving brand-safety assurance and enabling data-driven content strategy optimization.
  • Delivered measurement products identifying 400K+ violative accounts, enabling enforcement on 15K+ accounts without increasing appeals, and reducing manual review effort by 40%.
  • Secured dedicated compute capacity and senior stakeholder alignment enabling scalable processing and improving reliability by reducing query failures.
  • Leader for analytics as product initiatives, including resourcing for large initiatives, team allocation, hardware upgrades, and procurement of third-party software through enterprise compliance processes.
  • Led collaborative research partnerships with educational institutions and research organizations, leading the creation of an industry first external IRB methodology for ethical review and legal approval of research initiatives.
  • Deployed organization-wide automated detection systems with 95% accuracy, addressing highest-harm platform issues and protect brand safety and platform integrity.
  • Strengthened technical rigor and enforcement of consistent practices via mentorship and coaching on statistics, measurement, methodology, and research design.

Head of Accreditation

National Emergency Responder and Public Safety Center

Nov 2020 - Jul 2022

Led a team of 8 with end-to-end ownership of product development, market analysis, and accreditation compliance for programs serving 1,000+ agencies, with full P&L responsibility.

  • Drove B2B client acquisition by authoring winning RFPs, translating agency requirements into clear product scope and partnership strategy.
  • Owned vision and roadmap for analytics products, enabling data-driven operational decision workflows targeting retention and expansion.
  • Built strategic partnerships with industry-leading organizations to expand reach, strengthen credibility, and scale offerings.
  • Presented to government officials and oversight bodies, communicating risk and compliance considerations to guide leadership decisions in regulated environments.

Director of Evaluations and Analytics

Brower Psychological Police and Public Safety Services

Jul 2019 - Jul 2022

Directed three product teams (18 total members), owning delivery of digital psychological assessment products, people analytics platforms, and wellness programs in regulated public safety environments.

  • Increased operational throughput 5x by leading a digital transformation via cloud-based assessment products, while digitizing end-to-end workflows, improving speed, efficiency, and objectivity.
  • Secured $500K+ in grant funding by building data-driven roadmaps and demonstrating measurable DEI outcomes through proposals grounded in behavioral science and digital innovation.
  • Built partnerships with the national regulatory bodies, aligned offerings with regulatory expectations to achieve compliance, accreditation, and endorsements.
  • Won RFPs to expand a multi-state client portfolio, translating operational requirements into deliverable scope and driving revenue growth and market expansion.
  • Developed predictive analytics for mental health service forecasting, applying evidence-based measurement to guide operational improvements to identify and serve the employees of all client agencies fairly and equitably.

Doctoral Researcher & Supervisor

Fielding Graduate University

May 2015 - Jul 2019

Led supervision and training for 12 master's-level clinicians and practicum students across clinical and public safety settings, while conducting research in predictive behavioral analytics and clinically grounded measurement.

  • Diagnostic Psychological Evaluations: Cedar Springs Psychiatric Hospital.
  • Forensic, Occupational, and Clinical Neuropsychological Assessment: Rocky Mountain Behavioral Health.
  • Critical Incident Response, Fitness for Duty Evaluations, Psychotherapy: Brower Psychological Police and Public Safety Psychology.
  • Assistant Disaster Coordinator: Comprehensive Clinical Services (Aurora Mental Health Center).
  • Teaching Assistant: Theories of Psychotherapy.
  • Research Assistant: Clinically Predictive Linguistic Analysis of Thematic Apperception Test Narratives.

Staff Data Scientist

QMS Infotech, Inc.

Mar 2008 - Oct 2019

Developed advanced analytics products to audit private-sector hiring practices, identifying systemic bias and establishing psychometric and statistical standards grounded in regulatory compliance for B2B clients.

  • Delivered defensible analytics and audit frameworks enabling HR partners to implement bias-mitigating policies at scale while maintaining psychometric validity and legal compliance.
  • Established repeatable statistical modeling practices for audit workflows, producing auditable analyses and actionable recommendations that improved client hiring strategies and diversity outcomes.
Core Competencies

Technical & Leadership Capabilities

Leadership & AI Strategy
Technical LeadershipAI StrategyML LeadershipProduct StrategyCross-Functional LeadershipMentoring & CoachingOKR TaxonomyAI GovernanceResponsible AIRegulatory ComplianceBias MitigationEthical AI
Machine Learning & Modeling
Neuro-Symbolic AIGraph MLKnowledge GraphsEntity ResolutionWeak SupervisionReinforcement LearningAgentic SystemsLLMsNLPSemantic ClusteringPredictive ModelingXGBoostCausal InferenceAblation Studies
MLOps & Infrastructure
End-to-End Pipeline ArchitectureModel Lifecycle MgmtDistributed ComputingAWS SageMaker / EMR / EC2 / GlueDatabricksDelta LakeApache SparkAirflowDocker & KubernetesCI/CD
Research & Measurement
Experimental DesignA/B TestingPsychometricsSource Independence AnalysisBehavioral TypologiesIncrementality TestingMedia Mix ModelingLift Studies
Technical Stack & Tools
PythonPandas / NumPyScikit-learnPyTorchTensorFlowHugging FacePySparkSQL / Hive / ScalaKafkaTableau / Power BI