Namrata Sanyal

Developed and optimized Generative AI algorithms using LLMs and vector databases for Question Answering and network attack analysis via RAG pipeline and LangChain. Engineered a Deep Learning pipeline for Language Identification, Transcription, and Translation across 99+ languages with 98.4% accuracy, reducing processing time by 85%. Engineered a system to monitor malicious data transfers from sensitive email domains using Scala and Spark, analyzing 100,000+ email sessions daily. Designed a Deep Neural Network with Isolation Forest for Network Intrusion Detection handling 10Gbps raw traffic. Developed NLP-based algorithm for high-volume email text classification for 7 classes.

Speaker at Google I/O extended 2023. Conducted Machine Learning Study Jams on behalf of Google Developers Student Club. Organized workshops on Machine Learning. Took classes for over 100 students to guide them in Kaggle certification courses.

Extracted PII from text streams using NLP algorithms and ML models in Java. Worked on User Action Analysis based on classification of encrypted metadata. Worked on Indian Grid Reference System by reverse engineering APKs for coordinate conversion. Worked on extracting images from packet captures of Group 3 and Group 4 facsimile data.

Worked in collaboration with Dr. Subhamoy Maitra (ISI Kolkata) and Dr. Soumik Das on image processing and cryptanalysis of two DCT-based invisible watermarking schemes. Paper presented at the 3rd International Conference on Advanced Computing and Applications-2024.