Amin Salighehdar

About Amin Salighehdar

AI/ML expert and data science leader with a Ph.D. in Financial Engineering and 10+ years of experience delivering measurable business impact in fraud detection, credit risk, loss forecasting, and customer analytics.Proven track record in developing personalized recommendation systems using ML and Generative AI (LLMs) to increase conversion rates and optimize marketing initiatives. Experienced in designing and deploying experimentation frameworks including causal inference and A/B testing to measure model and product impact.Hands-on expertise in fine-tuning LLMs and transformers (BERT, GPT, LLaMA) with Hugging Face and PyTorch, and building RAG pipelines to scale enterprise NLP applications that improve engagement, personalization, and decision-making.Proficient in Python, SQL, PySpark, AWS (SageMaker), Hive, Hadoop, Git, and deep learning frameworks. Skilled in advanced ML techniques such as LightGBM, XGBoost, LSTM, clustering, identity graphs, and time series modeling.

Experience

1. Manager Data Science

   TikTok

   Jan 2024 — Present · San Jose, CA, US

   Designed and deployed real-time AI/ML fraud detection strategies in the payment and trust & safety domain to combat account takeover (ATO), bot attacks, and stolen card abuse, safeguarding millions of daily transactions and reducing capital loss rate by 50% through enhanced precision and adaptive thresholding.• Built behavior-based user segmentation and clustering to optimize fraud decisioning logic, reducing false positives and user block rates by 40% while sustaining strong defenses against ATO and synthetic identity fraud.• Built and executed A/B testing and experimentation pipelines to measure the impact of model updates on user behavior, conversion rates, and fraud outcomes.• Launched an incentive campaign that boosted transaction volume by 20%, while embedding fraud-risk safeguards to limit exposure to payment fraud and promotional abuse, protecting revenue and reinforcing platform trust.• Improved conversion rates by leveraging behavioral segmentation to develop personalized payment method recommendations, enabling frictionless user journeys and higher checkout completion.• Led a team of data scientists and fraud analysts to design AI/ML models that increased detection accuracy across ATO, card testing, and bot-driven fraud, significantly cutting false positives in payments.• Drove cross-functional collaboration with analytics, engineering, and product teams to deploy scalable AI-driven fraud systems aligned to KPIs, regulatory thresholds, and evolving fraud patterns, ensuring resilience against emerging adversarial attacks.

Education

• Stevens Institute of Technology

  Master's degree, Financial Engineering

  2014 — 2016

• Stevens Institute of Technology

  Doctor of Philosophy (Ph.D.), Financial Engineering

  2014 — 2018

Skills

• Data Analysis
• Programming
• Optimization
• Microsoft Office
• Sas Programming
• Java
• Statistics
• Autocad
• Matlab
• Mathematica
• Microcontrollers
• System Identification
• Algorithms
• Image Processing
• C++
• Machine Learning
• Research
• Data Mining
• R
• Labview
• Linux System Administration
• Python
• Latex
• Sql
• Microsoft Excel
• Digital Signal Processors
• Mathematical Modeling
• Anova