Ananya Joshi

Automated and deployed an ML-driven fraud detection framework using unsupervised anomaly detection with AWS Lambda and S3, achieving 100% detection precision and <4.5% false positives on validated cases. Implemented a credential stuffing detection model using multi-temporal features, weighted velocity modeling, and ensemble machine learning, achieving 90.54% precision and 97.27% recall.

Built AI-powered microservices for a skills-first hiring platform using Azure OpenAI and FastAPI, including a resume intelligence module and an NLP-based skills extraction engine. Developed an adaptive assessment engine pairing LLM-generated questions with an ML-driven model calibration for personalized candidate evaluation.

Designed a credit risk modeling framework across multiple customer segments, using statistical feature selection and model benchmarking to reduce 200+ variables to 10 key predictors. Improved credit risk models with feature engineering, ensemble methods, algorithm evaluation, and hyperparameter optimization, increasing the Gini coefficient by 21% and improving predictive risk discrimination.

Improved sentiment analytics pipelines for social media data using text preprocessing and feature extraction to uncover engagement trends. Created Power BI visualizations for real-time insight generation and performance monitoring.

Grade: CGPA: 3.8 / 4.00

Pursuing a B.Tech in Data Science at NMIMS MPSTME (Mumbai), with a CGPA of 3.8/4.00.