Sagar Raichandani Email and Phone Number

Grad TA for a math-focused Machine Learning course (CSCI 1952Q) offered by the Department of Computer Science at Brown University. My key responsibilities include: conducting review sessions of the course material, holding office hours to help students, and grading student assignments & projects.

Head TA for the Machine Learning Theory course (DATA 2060) offered by the Data Science Institute at Brown University. My key responsibilities include: conducting review sessions of the course material, developing homework components, holding office hours to help students, and grading student assignments & projects.

• Engineered critical features to detect fraudulent cancellations on the Uber Eats platform. Mitigated fraud revenue loss of over 15 basis points through data-driven insights, fraud rules, and machine learning models.• Gathered requirements from DS and ML Engg. org and implemented redesign of user-risk segmentation; enhanced differentiability of risk sub-segments by more than 50% through features quantifying suspicious device, payment, and engagement behavior.

BlackBuck is a digital marketplace for trucking logistics in India. I was responsible for identifying and analyzing user challenges through data, model relevant solutions, and monitor impact post feature launch. Made foundational contributions to problems in Risk & Fraud and Pricing:• Developed anomaly detection models in Python to identify fraudulent users and fake load posts; built signal scores that identified fraudulent users based on in‑app behavior, geo‑dispersion, point‑pattern randomness, and K‑means clustering.• Built the customer risk‑categorization lifecycle system, automated ETL workflows to move data from OLTP databases to centralized data lake on S3, and facilitated actionable reports for the operations team through daily crons.• Conducted root-cause analysis to identify product exploits; worked with Business, Product, and Engineering stakeholders on prioritization and validation of fixed.

• Applied DBSCAN to identify demand‐supply clusters; built a time-series model that predicted daily prices across these clusters, thereby maintaining profitability on operating routes.• Fulfilled end‐to‐end responsibilities; deployed model to production, maintained logs in MySQL databases, and integrated with engineering services through APIs to reduce manual intervention and operational error.

Unbxd is an e-commerce search provider for 1300+ global online brands. As a Business Analyst, I played an integral role in: designing A/B tests to guage performance of ML models; analyzing terabyte scale data for insights on product features; and developing critical pipelines for automated internal & external reports.• Built a PySpark framework to identify descriptors (attributes) in JSON product catalogs to be indexed by Solr backend systems. In addition to increasing relevance F-score, the framework saved 100+ QA man-hours each month.• Created a Python package using G-Suite APIs to enable automation of data pipelines and reporting in Google BigQuery, Google Analytics, Google Sheets and Gmail.• Analyzed search data, performed A/B tests, and built SQL reports to provide insights on product features, search relevance, and ranking models that boosted conversion by 8-12%.