Shramita Maheshwari

I’m an AI Intern focused on production-ready machine learning and GenAI systems, with hands-on experience building end-to-end pipelines and shipping solutions that reduce real operational effort. At Infosys Springboard, I developed a Real-Time Competitor Strategy Tracker that processes pricing and inventory signals across 250+ products, cutting analyst workload from 20 to 4 hours/week (80% reduction).

My work centers on NLP and reliability: I’ve applied transformer-based techniques, engineered scalable data workflows, and specifically investigated LLM hallucination failure modes to improve relevance and reliability through prompt design and evaluation. I also built a hallucination-resistant Conversational Assessment Recommender using a production RAG architecture (LangChain, SentenceTransformers, FastAPI) and designed vector search APIs with systematic chunking/embedding benchmarking (FAISS), along with strong ML foundations from course work and projects including forecasting and CNN classification.