Shweta Shardul

WORK HISTORY

EDUCATION

ABOUT SHWETA SHARDUL

Fuel me with curiosity and caffeine ~ in 14 days, I’ll not just learn it, I’ll ship it. Recently, I’ve been diving deep into LLMs, RAG pipelines, and prompt engineering, experimenting with content generation, workflow automation, summarization, and data extraction. I’ve built working prototypes that connect real-world data with generative AI, making insights more accessible and actionable. These hands-on projects taught me how to reduce hallucinations, optimize prompts, and design practical RAG workflows.🧠 Graduate Researcher at NJIT, executed 500+ experiments across CNNs, Transformers, and MLPs to study Neural Collapse, improving feature convergence by 25% and cutting test error by 18%.🤖 At TCS, built and scaled intelligent chatbot systems for a global pharmaceutical client, boosting NLP model accuracy from 46% → 83%, while also designing Python-based test automation and CI-backed validation pipelines to ensure reliability, regression safety, and production correctness across high-traffic deployments. Designed MLOps pipelines on AWS MWAA & Python, scaling ETL ingestion from 70K to 800K+ records with 90% effort reduction, ensuring reliable CI/CD deployments.‍ Teaching Assistant for Data Science, mentored students on ML algorithms, evaluation techniques, and implementation with Scikit-learn, TensorFlow, and Python. President of Graduate Women in Computing Society (GWiCS), leading a 200+ member community and increasing engagement by 170% through tech talks, mentorship, and conference representation at GHC, SWE, and SHPE.Forget the buzzwords, throw me into something new, and I’ll figure it out fast. I learn deeply, move quickly, and make people feel at ease while getting things done.