Samrat Halder Email & Phone Number

Seattle, WA, US

Traveler Voice and Content (LLMs, GenAI in Recs)

Seattle, WA, US

Marketplace Trust & Safety, Squad Lead

Seattle, WA, US

• Fraud & Risk
• Booking
• Supply

New York, NY, US

We find in our multi-tasking neural model for learning psycholinguistic aspect of language, SRLs play a big part; surprisingly, random embeddings with tuning do as good as GloVe, and more data do not necessarily yield better results. We publish our new SotA results on thematic fit tasks at BlackBoxNLP (EMNLP 2022).

Newark, NJ, US

Built a customer targeting model combining multi-arm bandit and xgboost which resulted in 0.8% improvement in CTR for email campaign of Audible plus and premium

Newark, NJ, US

• Did research on data science models used by practitioners for marketing campaigns to optimize customer targeting, and acquisition leveraging gradient boosting trees, lift model, contextual bandit
• Tech-stack used include AWS (Sagemaker, Step Function, s3), Alation, PySpark, SQL

New York, NY, US

Worked with Dr. Simona Abis on multiple projects that involved applied ML, and NLP to understand and quantify the impact of Artificial Intelligence in the changing economics of knowledge production

London, -, GB

• Collaborated on an open-source data science project to analyze various virus and economic models and come up with possible country-specific lockdown exit strategies post COVID-19 pandemic that minimizes the impact of.
• Research Area: Disease models, multi-objective optimization

Liaison between Engineering Graduate Student Council (EGSC) and Data Science Institute Student Council (DSISC)

New York, NY, US

Led weekly programming classes, held recitation sessions, and graded assignments for a graduate-level Python programming course

London, Middlesex, GB

• Supervised a team of analysts based in New Delhi, IN
• Spearheaded technical and analytic support to 2 clients in India for end-to-end integration of API, automated order execution, and live fund management worth USD 1M

London, Middlesex, GB

• Advisor and CEO: Professor Gautam Mitra (https://en.m.wikipedia.org/wiki/Gautam_Mitra)
• Worked on multiple R&D projects which include machine learning, large-scale portfolio analytics which improved volatility and return on our index funds
• Managed a summer intern for enhancing trade signals by including news sentiment-based signals

London, Greater London, GB

Workshop on stochastic optimization and integer programing organized by OptiRisk Systems and AMPLDev

Mumbai, Maharashtra, IN

• Developed POCs on AWS (Amazon Web Services- Kinesis, Lambda, SQS, Cloudwatch, Dynamo DB), End to end automation of several tasks, Data cleaning
• Research Area: Volatility smile, Option pricing, Deep learning

Berlin, DE

• Built a model using LDA (topic modeling) and sentiment analysis with Tweets on interest rates to get Hawkish- Dovish sentiment about the market
• Built a Structural Vector Autoregression model to estimate the impact of market sentiment on asset prices
• Developed a Shiny Dashboard to monitor near real-time sentiment on political parties from rss feeds

• Selected among the top 6 investment teams among 24 teams in the final round
• Presented our valuation of Reliance Industries based on dividend discount and price multiple methods at the Universitas of Indonesia, Jakarta

• Advisor: Dr. Shravan Hanasoge
• Published paper: Finite-frequency Sensitivity Kernels in Spherical Geometry for Time–Distance Helioseismology
• Research Area: Computational astrophysics & astronomy, Numerical Methods, Parallel Programming, Solar Kernels, Time-Distance Helioseismology, Boundary value problems, PDE's

Mumbai, Maharashtra, IN

National Initiative on Undergraduate Science (NIUS) Camp - Selected among 45 freshmen students in Science and Engineering across the country for a summer camp on advanced scientific computing

Master Of Science - Ms, Data Science

Bachelor Of Science (Honours) And Master Of Science, Physics, Mathematics, Financial Engineering