Md Ehtesham

Spearheaded AI-driven product enhancements and collaborated on innovative solutions to improve operational efficiency and revenue growth for a travel technology company.

• Spearheaded performance enhancements of an AI-driven voice call bot and fine-tuned LLM prompts, boosting qualified lead identification accuracy from 40% to 75% and reducing the need for manual pre-seller intervention.
• Improved AI Instagram DM bot's performance, increasing customer satisfaction rate from 20% to approximately 50% through strategic optimization and feature implementation.
• Collaborated on the end-to-end development of an AI-powered travel itinerary builder for the company's new website, enabling real-time, budget-based trip planning.
• Implemented automated customer support features within the itinerary builder, directly contributing to a 5% increase in company revenue.

Engineered and localized a Retrieval-Augmented Generation pipeline to enhance AI teaching assistant accuracy and accessibility.

• Engineered a Retrieval-Augmented Generation (RAG) pipeline, leveraging OpenAI embedding models and FAISS vector store, significantly improving AI teaching assistant's contextual accuracy by 30% (measured by BLEU score).
• Localized the AI assistant for a regional language, enhancing accessibility for diverse educational use cases and expanding user reach.
• Developed robust data processing and integration strategies for PDFMiner, ensuring efficient extraction and utilization of unstructured text data within the RAG pipeline.
• Utilized Langchain framework to streamline the development and deployment of LLM-powered applications, accelerating project timelines.

Conducted research and developed a machine learning system for early lung cancer detection using sensor data and IoT.

• Collected and meticulously processed sensor data from 1000+ samples from Organic Electrochemical Transistor (OECT)-based devices to build a comprehensive dataset for predicting potential cancer infections in biological samples.
• Applied advanced Multinomial Logistic Regression for multi-class classification, achieving a high prediction accuracy of up to 98% in complex cancer detection tasks.
• Designed and implemented the system using a combination of Machine Learning and IoT technologies, demonstrating expertise in interdisciplinary research and application.
• Contributed to a critical public health initiative by developing a predictive model that could aid in early disease detection, showcasing problem-solving skills in a high-impact domain.

Developed a deep learning model to analyze and classify COVID-19 cases from chest X-ray images.

Developed a full-stack online forum platform for the IITJ community to facilitate knowledge sharing and discussions.

Ranked in the top 2% with an A grade in Data Structures, demonstrating strong foundational computer science knowledge.

Secured a rank in the top 1% among 10 lakh+ candidates in the highly competitive JEE 2021 examination.

Secured 1st position in Kanpur city with 97.8% marks in the XII Boards examination.

Qualified Level 1 and Level 2 International Mathematical Olympiad (IMO) exams multiple times, demonstrating exceptional mathematical aptitude during school years.