Yash Aggarwal
About
Detail
AI/ML Engineer @ LivePerson | Ex-ZS Data Scientist | Building Autonomous Agents & RAG Pipelines | Delivered $20M+ Business Impact
National Capital Territory of Delhi, India
Résumé
Jobs
verified_user
0% verified
Ml/AI InternLivepersonJun 2025 - Current (1 year 5 months)- Research Assistant Cognitive AI LabIndiana University BloomingtonSep 2024 - Current (2 years 2 months)As a Research Assistant in the Cognitive AI Lab, I enhance AI models by applying insights from human brain cognition, exploring parallels between neural mechanisms and deep learning. I collaborate with a multidisciplinary team to advance AI research and its real-world applications.
- Research Assistant- Biomedical AI LabIndiana University BloomingtonSep 2024 - Current (2 years 2 months)As a member of the Biomedical AI Lab, I focus on advancing methodologies that leverage Natural Language Processing (NLP) and Large Language Models (LLMs) to transform and enhance biomedical data. My work involves developing cutting-edge techniques to extract valuable, structured information from vast and complex datasets, streamlining this data for use in critical research and discovery efforts within the biomedical field.
- Advance Data Science Associate ConsultantZSJul 2022 - Aug 2024 (2 years 2 months)As a Data Science Associate Consultant at ZS, I designed tailored machine learning and artificial intelligence solutions to meet specific client needs. My role encompassed direct interactions with clients to understand their objectives and challenges, as well as leading and managing a team to ensure the successful deployment and integration of these solutions. • Designed and implemented AI-driven solutions using GenAI, RAG, and LLMs, addressing key challenges for pharmaceutical R&D clients and enhancing operational efficiency. • Led the development of an automated knowledge extraction tool, reducing manual review time by 40% and saving 200 hours annually for clients. • Developed a high-precision entity extraction pipeline for clinical proto
- Advance Data Science AssociateZSJul 2020 - Jul 2022 (2 years 1 month)In my role as a Data Science Associate at ZS, I focused on implementing tailored machine learning and artificial intelligence solutions efficiently. My responsibilities included optimizing these solutions to meet specific client objectives and ensuring their successful deployment and integration, directly contributing to enhanced client operations and decision-making processes. • Developed and implemented ML solutions to optimize clinical trials, increasing efficiency by 25% and reducing setup times by 15% for multiple clients saving over 20 million dollars in operational cost. • Delivered predictive insights on protocol amendments, leading to a 20% reduction in unnecessary amendments and significant cost savings. • Predicted NASH risk with
- Machine Learning Teaching AssistantCoding BlocksSep 2018 - Mar 2019 (7 months)As a Machine Learning Teaching Assistant at Coding Blocks, I was responsible for elucidating complex machine learning concepts in an accessible and understandable manner for students. My role involved designing and delivering instructional content, creating assignments, and developing practical tasks that facilitated a deeper understanding of machine learning applications. This position required me to not only impart knowledge but also ensure that students could effectively implement these concepts in real-world scenarios.
Education
verified_user
0% verified
- Master's degree, Data ScienceIndiana University BloomingtonJan 2024 - Jan 2026 (2 years 1 month)
- Bachelor of Technology, Information TechnologyMaharaja Agrasen Institute of TechnologyJan 2016 - Dec 2020 (5 years)
Projects (professional or personal)
verified_user
0% verified
- JJanssen Protocol DigitizationNov 2023 - Aug 2024 (10 months)• Role: Led the design and implementation of a high-precision entity extraction pipeline from clinical protocol documents. • Impact: Streamlined the entity extraction from clinical protocols by automating the process, achieving a threefold increase in efficiency. This automation also enhanced training models and pipelines, enabling the optimization of protocols through insights derived from historical data.
- KKnowledge Engineering PlatformJan 2023 - Dec 2023 (1 year)• Role: Spearheaded the development of an automated knowledge extraction tool using LLMs from PubMed medical articles, integrating a chatbot for efficient evidence gathering. • Impact: Reduced manual article review time by 40%, enabling clients to redirect approximately 200 hours annually toward higher-value research tasks.
- JJanssen Layman SummaryDec 2022 - Feb 2023 (3 months)• Role: Managed the creation of a summarization pipeline for patient lay summaries from clinical protocols. • Impact: Boosted client document processing efficiency by 50%, facilitating faster patient communications and increased summary production from 20 to 30 summaries per month
- BBMS Amendment Extraction And PredictionJan 2021 - Aug 2021 (8 months)• Role: Oversaw the analysis and systematization of over 10,000 protocol amendments to develop predictive models. • Impact: Delivered actionable insights into amendment characteristics, empowering clients to reduce unnecessary amendments and achieve cost savings of up to 20% in trial operations.
- CCTFO and CTEMPJul 2020 - Dec 2020 (6 months)• Role: Directed the development of a clinical trial optimization tool for site selection and patient enrollment forecasting. • Impact: Enhanced clinical trial efficiency by 25%, leading to a 15% reduction in trial setup times for major pharmaceutical clients like Sanofi, J&J, and BMS.
- IImage CaptioningFeb 2019 - Mar 2019 (2 months)Created a model which given the image generated the appropriate caption for the image using combination of Lstm and inception model.
- BBlack Jack AiJan 2019Created a reinforcement learning model to teach my computer to play blackjack.
- TTaught Quadcopter how to fly using reinforcement learningJan 2019 - Feb 2019 (2 months)
- FFace RecognitionCreated a Real time face detector in a video using opencv and Haar cascade.
- MMovie Recommendation based on Item based collaborative filteringCreated a moving Recommedation system using Pandas Library Which Given the movie u are interested in or liked recommend similar movies.
- HHouse Loan Riskparticipated in kaggles House Risk Prediction challenge in which i had to analyse massive real world data and create a machine learning model for the same.
- DDog Breed ClassificationCreated a cnn model with the help of transfer learning which was able to predict breed of the dog given the image with ~80% accuracy.
- CCustomer SegmentationCreated customer Segmentation using unsupervised Learning methods like Gaussian mixture and Kmeans.
Awards
verified_user
0% verified
- CWinner Fall Narrative ChallengeCenter of Disease ControlDec 2023Led a team to victory in this hackathon hosted by the CDC, developing unsupervised learning techniques to extract and analyze fall incident narratives for elderly patients, and improving preventative care strategies.
- GWinner Smart India Hackathon(2019)Goverment of India and Sun pharmaceuticalsMar 2019Created a deep learning model which will take problem encountered in factory or industry as input and then produce root cause analysis for that problem.I also created some additional analytical tools to help them understand the root cause more accurately.I also did extensive data analysis to find out major root causes and derive relevant information from the data provided.