Lead Machine Learning Engineer · TikTok U.S. Data Security · New York

Finding the one bad actor among 150 million good ones.

I build machine learning systems for fraud, abuse and account security — from feature design and label governance to low-latency models running in production against adversaries who adapt every day.

Get in touch View résumé
Years applying ML to fraud detection & account abuse
9+
Daily active users protected from Account Takeover (ATO) risk
150M+
ATO perception after applying sequence-based behavioral modeling
5→60%
Recall in stopping surgical zero-click attacks on high-profile news & celebrity accounts
95%
01About
Darshan Patel smiling on a New York City street

Hi, I’m Darshan. I lead the Account Security machine learning team at TikTok U.S. Data Security, where we protect more than 150 million people a day from account takeover.

I’ve spent over nine years applying machine learning to fight adversarial forces and defending modern businesses. I started at PwC building fraud gradient-boosted models and targeted credit card fraud risk strategies for financial institutions, then spent three years at DataVisor using unsupervised and graph-based machine learning for some of the world’s largest banks and e-commerce platforms. Today I work across the full model lifecycle: sequence models, gradient-boosted trees, graph analytics and anomaly detection, all designed to keep learning as attackers change tactics.

I earned a B.S. in Computer Engineering cum laude from Penn State and an M.S. in Computer Science, specializing in machine learning, from Georgia Tech. Outside work you’ll find me running, reading, traveling, or adding to a collection of travel mugs that is getting out of hand.

02Experience

  1. Dec 2023 — Present

    New York, NY

    TikTok U.S. Data Security

    Lead Machine Learning Engineer, Account Security

    • Lead the team responsible for Account Takeover (ATO) risk control across 150M+ daily active users. Designed the architecture that organizes every model and strategy into three pillars (defense, perception and label governance) connected by a feedback loop that keeps us ahead of adversaries.
    • Designed, launched and now monitor the ATO user-journey sequence model, raising ATO perception from 5% to 60%.
    • Uncovered and dismantled a hacker network running targeted attacks on high-profile news and celebrity accounts, reaching 95% recall.
    • Cut phishing prevalence 15% QoQ with low-latency CatBoost defenses, and reduced ATO on political and high-value accounts by 40% during the 2024 U.S. election.
    • Grew high-value account security coverage from 1% to 5% of MAU. The segmentation now runs on every TikTok user engagement surface.

    Recognition TikTok Hackathon winner (RiskAI Copilot, now in production) · Two “Exceeded Expectations” performance ratings

  2. Aug 2020 — Dec 2023

    Mountain View, CA

    DataVisor

    Senior Data Scientist, Fraud Detection & Research

    • Tackled payments fraud, application fraud, account takeover and money laundering with <1% false-positive rate and 70%+ recall.
    • Led projects and mentored junior data scientists on graph analytics for semi-labeled problems and identity graphs for entity resolution at massive scale.
    • Improved DataVisor’s proprietary unsupervised models for some of the world’s largest financial and e-commerce platforms.
    • Supported go-to-market with a 50% lift in deal conversion, contributing to $1.2M+ in new business.

  3. Feb 2018 — Aug 2020

    Philadelphia → New York

    PwC

    Data Scientist, Fraud Detection · Senior Associate

    • Built, monitored and improved fraud strategies for financial-services clients using big-data tools and machine learning.
    • Led a team of backend, UI and ML engineers across the US, Europe, India and China to build PwC’s integrated fraud-risk analytics solution.
    • Designed acquisition, lifecycle, fraud and operations campaigns that delivered $10M+ in top-line savings.

Earlier

  • PwCData Science Intern, Forensics Technology Services2017
  • ScratchworkiOS Developer & Project Lead (Senior Design)2017
  • Penn StateLearning Assistant, Data Structures & Algorithms2017
  • OracleSoftware Engineering Co-op, Primavera Prime2016
03Expertise

Machine learning

  • Sequence models
  • Gradient-boosted trees (CatBoost)
  • Graph analytics
  • Entity resolution
  • Unsupervised detection
  • LLM applications

Languages & frameworks

  • Python
  • SQL
  • PyTorch
  • TensorFlow
  • Spark
  • Java
  • Go
  • C
  • NetworkX
  • JavaScript / Node.js
  • Swift

Data & platforms

  • Spark MLlib
  • Hadoop
  • scikit-learn
  • PostgreSQL
  • Tableau
  • Power BI
  • Alteryx
  • AWS EC2 / EMR
  • GCP Kubernetes

M.S. Computer Science · 2023

Georgia Institute of Technology

Specialization in Machine Learning (OMSCS). Coursework includes Machine Learning, Deep Learning, Reinforcement Learning, Artificial Intelligence, ML for Trading, Data & Visual Analytics, Graduate Algorithms and Database Systems.

B.S. Computer Engineering · 2017

Penn State University

Cum laude, Dean’s List every semester, 3.84 GPA. Tau Beta Pi, HackPSU, TEDxPSU, and Vice President of the Robotics Club.

Honors

Recognition

  • TikTok Hackathon Winner2024
  • Dr. Ayoub Mathematics Award2015
  • President’s Freshman Award2014
04Side projects

Projects I built for fun and for learning. The code for each is on GitHub.

05By the numbers

On the job

Data science client projects
31
Solutions engineering client projects
110
Personal projects
10

Off the clock

Travel mugs collected
45+
Books read in 2026
9
Miles run since 2020
1,074
06Contact

Working on trust & safety, fraud, or security ML? Let’s talk.

darshanpatel25894@gmail.com