J
Jhanavi Putcha
About
Detail
New York, United States
Résumé
Jobs
verified_user
0% verified
- TAI/ML EngineerTriosoft LLCMay 2025 - Current (1 year 1 month)• Constructed transformer-based NLP models for automated extraction of financial insights from 87+ unstructured reports per week, reducing analyst effort by 18% and improving signal-to-noise ratio in investment research pipelines. • Engineered reinforcement learning models for portfolio optimization, simulating 24+ market scenarios per asset class, achieving a 6% improvement in risk-adjusted returns and enabling 95+ scenario analyses for senior investment teams. • Implemented PyTorch/TensorFlow deep learning models for credit risk scoring and market anomaly detection, processing >15 million transaction records monthly, achieving 93% coverage of critical high-risk events. • Deployed generative Al models and LLMs to automatically generate inv
Machine Learning ScientistLTIMindtreeJan 2022 - Jul 2024 (2 years 7 months)• Architected time-series forecasting models on Spark and Snowflake for predicting vehicle component failures from IoT sensor data, improving predictive maintenance accuracy by 18% and reducing unplanned downtime by 12% across 25+ production lines. • Engineered real-time anomaly detection pipelines using LSTM and XGBoost on sensor and telemetry data streams, identifying 83 critical anomalies per month and increasing system reliability by 15%. • Designed end-to-end ETL pipelines in Python and PySpark to process 20 TB/month of vehicle data, ensuring low-latency (<5 sec) ingestion and enabling 95% data availability for ML model training. • Developed interactive dashboards using Power Bl, providing 95+ actionable KPIs across fleet health, predi
Education
verified_user
0% verified
- Master of Science in Artificial IntelligenceUniversity at BuffaloBuffalo, New York
Projects (professional or personal)
verified_user
0% verified
- AAI-Powered Credit Risk Assessment System• Engineered a multi-class classification system to categorize students into performance tiers, combining historical grades, participation, and behavioral patterns, achieving 87% prediction accuracy and enabling targeted academic support programs. • Performed feature selection and dimensionality reduction using PCA and correlation analysis, improving model efficiency and interpretability while maintaining 87% prediction accuracy for credit risk classification.
- IIntelligent Student Performance Predictor with Early Intervention System• Built predictive models using Random Forest, XGBoost, and LSTM to forecast student academic performance with 85-90% accuracy, leveraging 20+ engineered features from attendance, assignments, and engagement metrics, enabling early identification of at-risk students. • Developed an interactive Flask-based dashboard with SHAP explainability to visualize predictions and feature impact, allowing educators to understand key performance drivers and intervene 2-3 weeks before critical exams, improving actionable insights and decision-making.