AI/ML engineer with 3+ years of experience in building accurate and reliable NLP models and RAG systems. Efficient in multi-stakeholder management, disambiguating complex problem statements, processing data, generating insights, building models, and deploying them into production. Adept at using Python, LangChain, Streamlit, PyTorch, LLMs, Transformers, Hugging Face, Git, SQL and AWS.
Experience
Data Scientist II
Allen Digital
Developed and deployed an end-to-end gRPC service to automate student doubt resolution using a robust multi-agent Retrieval-Augmented Generation (RAG) framework, achieving 98% accuracy, a 12% improvement over the baseline. Enhanced user experience by integrating image-based output capabilities and a fine-tuned audio model. Designed and implemented safety features for content moderation and jailbreak prevention. Deployed LLMs on Amazon SageMaker using the low-latency Text Generation Inference (TGI) framework.
Data Scientist
Amazon
Developed a supervised learning solution to detect male bias in Amazon Job Descriptions (JDs) using a novel causal inferencing approach. Used BERT and Random Forest Regressor, adding an explainability layer with KernelSHAP. Built an Amazon Gift Cards review classifier with 78% accuracy, deploying the end-to-end model pipeline using AWS tools. Developed an unsupervised learning solution for identifying themes in employee responses using Sentence Transformers and hierarchical clustering.
Software Engineer Intern
Avaamo
Designed and built conversational AI chat bots using IBM Watson and Google DialogFlow, implementing intelligence features like episodic memory and resume conversation.
Product Intern
Adobe
Performed document generation by replacing text tags in input word or PDF templates with input data.
Education
BITS Pilani
B.E.
Computer Science
BITS Pilani
M.Sc
Mathematics