Pradyumn Pathak
Masters in AI | AI Engineer | Gen-AI Researcher | 3+ years of Industry AI Experience. Futurist looking ahead in time to see how AI and Machine Learning will open new possibilities for mankind. @DePaul, ex @FPT-AI
- Role
- Gen-ai Research Assistant at Depaul University Jarvis College Of Computing And Digital Media
- Location
- Chicago, IL, US
- LinkedIn followers
- 500 followers
About Pradyumn Pathak
AI Engineer and Researcher with 3+ years of experience developing and deploying intelligent systems across healthcare, education, and retail. Known for transforming innovative ideas like whiteboard digitization and real-time food consistency tracking into edge-ready, production-grade solutions. Adept at scaling deep learning models, optimizing performance on constrained hardware, and pioneering new approaches in NLP research. Recognized by Upsilon Pi Epsilon for academic excellence and contribution to computing. I also hold a bachelor\'s in Electronics Engineer degree from University of Mumbai. My passion towards AI comes from my time there working in robotics where I represented my university in national level robotics competition E-YIC. Naturally born with a passion for engineering, and problem solving. A critical and creative thinker, looking to always pushing the boundaries and always down to collaboration! Some of my hobbies are playing Dungeons and Dragons, soccer, video games, and working on creative projects ideas!
Experience
Gen-ai Research Assistant
Depaul University Jarvis College Of Computing And Digital Media
Sep 2024 — Present · Chicago, IL, US
Contributing to cutting-edge research in Generative AI, focused on building LLM-powered scientific discovery tools. My work involves building and leveraging these tools to capture scientific knowledge evolution across time, with the long-term vision of predicting emerging scientific breakthroughs.• Designed a novel Hierarchical Embedding strategy using \"Context Trees\" to extract stable, semantically rich representations of scientific keywords, reducing drift in LLM embedding space across temporal slices of scientific literature.• Built automated scripts to ingest, sanitize, and preprocess a 2M+ research paper dataset from Semantic Scholar, improved data cleanliness and training stability for downstream Sci BERT model training.• Optimized Sci BERT training pipeline, reducing training time by ~75% through data batching and preprocessing enhancements, accelerating research iterations.
Education
Kendriya Vidyalaya
High School Diploma, Computer Science
DePaul University Jarvis College of Computing and Digital Media
Masters in Science , Artificial Intelligence
University of Mumbai
Bachelor of Electronics Engineering, Electrical and Electronics Engineering
2016 — 2020
Find verified contacts for anyone on LinkedIn
Unifers gives sales teams verified emails and direct dials, enriched profiles, and outreach that lands in the inbox.
Free plan included · No credit card required
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.