Nisha

About Nisha

Data Scientist with experience designing and scaling machine learning systems on large, real-world datasets, where model decisions directly influence product behaviour and business outcomes.My primary expertise is in applied machine learning — building end-to-end solutions that span data preparation, feature engineering, model development, evaluation, thresholding, and production pipelines. I have worked extensively on high-impact problems involving imbalanced classification, risk scoring, and decision systems operating at scale.Beyond classical ML, I have hands-on experience and strong conceptual understanding of neural network architectures, including CNNs, RNNs, LSTMs, GRUs, and Transformers, and I actively explore how these approaches can be applied to complex pattern-recognition and sequence-based problems.I find my interested in:• Solving ambiguous, high-impact product problems• Translating business goals into measurable ML objectives• Designing models and systems that are robust, interpretable, and production-ready• Working on challenges where data science can create meaningful, visible change in a product or platform

Experience

1. Data Scientist

   Straive

   Mar 2025 — Present · Hyderabad, IN

   Worked on large-scale card fraud detection systems across CAMS (Compromised Account Management System) and Enumeration Fraud for multiple regions transaction\'s data (North America, Latin America, Europe ).Built end-to-end ML pipelines covering data ingestion (Hive/Hadoop), feature engineering (velocity & behavioural signals), feature reduction (VarClus), model training (LightGBM, Random Forest with Optuna HPT), threshold optimization, and risk tiering under highly imbalanced data. Evaluated ensemble strategies (probability aggregation, voting, stacking exposure) to improve robustness, achieving ~2–3% uplift in fraud capture over existing baselines while controlling false positives.

Education

• Indian Institute of Technology, Mandi

  Bachelor of Technology, Iit Mandi

  2018 — 2022