
An AI-powered autonomous interview preparation platform
Will be joining the global analytics and reporting (GAR) team of D.E. Shaw India in the Human Captial Domain
• Orchestrated a 5-layer end-to-end interview automation platform, decreasing interview process times by over 50% • Implemented a robust RAG pipeline to retrieve relevant Q&A pairs from vector embeddings, providing ground truth for LLM evals • Developed a production-ready FastAPI backend with Poetry dependency management ensuring seamless CI/CD pipeline • Managed 7 tables across PostgreSQL using SQLAlchemy ORM and Alembic migrations for schema versioning • Enhanced candidate preparedness by providing 24/7 AI-powered mock interviews, leading to increased candidate confidence
• Engineered ML web servers, PKM2Pred and Amylo-IC50Pred, to accelerate drug discovery for cancer and Alzheimer's. • Built high-accuracy classification models to identify and categorize therapeutic compounds. • Developed robust regression models to predict compound potency and bioactivity (AC50/IC50). • Translated the predictive pipelines into publicly accessible web servers deployed over BIT Mesra's linux servers
• Helped in the development of the website for Bitotsav’25, the largest cultural college fest in Jharkhand, used by 6000+ people. • Engineered a chat bot having a median response time of 500ms by designing a custom data ETL pipeline • Performed server optimizations and introduced caching to avoid redundant calculations and reduce response times by 75%. • Resolved 300+ technical inquiries, resulting in a 40% increase in online engagement and a 27% increase in event participation.
• Pioneered the 3-phase development of a Industrial Power Utilization prediction model to reduce energy wastage. • Reduced RMSE by 79% through comprehensive data processing and exploratory analysis using Pandas/NumPy. • Achieved an R² score of 0.9995 by optimizing key features and hyperparameters ensuring robust performance. • Implemented efficient dockerization and performed stress-testing and deployment over a standalone k8s cluster.
/usr/bin/send-mail/to-arya