Jimit Mistry

About Jimit Mistry

Machine Learning Engineer with Research experience in NLP, Human-Computer Interaction and Computer Vision. Specialize and have experience in building end-to-end NLP, Deep Learning and Machine Learning products for SaaS environment following MLOps strategies. I build predictive data-intensive applications in a corporate environment in an individual as well as a team capacity. I aspire to make a name for myself in the industry by helping major industries achieve their goals by innovating solutions and working with teams across the US.What I bring with me are the research and technical skills for building end-to-end Machine Learning/Deep Learning solutions and the personal acumen to work closely with SMEs, understanding their problems, and work within team environments to take the problems to resolution. My long-term focus, excellent communication, leadership, and strong work ethics guide me in my pursuit of excellence.Skills: Data Science, NLP, Deep Learning, Machine Learning, Data Engineering, Data Analytics, Big Data Analytics, Data Warehouse, Business Intelligence

Experience

1. Senior Machine Learning Engineer

   Infinx

   Jul 2024 — Present · Austin, TX, US

   Building their most advanced LLM Agents and Copilots.First ML Engineer/Data Scientist hired after the Chief Data Scientist.Strategize to build AI Agents which actually work. Play a key role by guiding their current and next generation of AI products.Designed AI systems for- Clinical Guideline Verification AI; making LLMs solve for domains much harder for them natively. Robust Generalization- Document AI Agent with Automated dynamic model training for new clients; Robust approach, generalization- Analytics AI Agent; R&D, industry first, completely dynamic, privacy first designR & D- Custom GEPA implementation to avoid it getting stuck at local minima for single call LLM instances- AI Analytics LLM Agent: Reasoning Model and Standard Model pipeline, dynamic selection of graph based multi step execution or all in one execution. Domain Agnostic, Data Size Agnostic, Privacy first (Model never sees the actual data)- Knowledge/Model distillation from Vision LLMs to DINOv3 for CPU only Vision Capabilities- Generalized pdf parsing: solved for multi page style and long documents- Automatic Onboarding of new Data for AI based Products- LLM out of domain expertise: Utilized clever design, prompt tuning and fine-tuning approaches to ensure LLMs work with complex domains with high performance.

Education

• Syracuse University

  Master of Science - MS, Data Science

  2019 — 2021

• National Institute of Technology Surat

  Bachelor of Technology (BTech), Electrical, Electronics and Communications Engineering

  2014 — 2018

Skills

• Kaizen
• Big Data
• Javascript
• Continuous Improvement
• Core Java
• Machine Learning
• C
• Python
• 5s
• Artificial Neural Networks
• Process Improvement
• Computer Vision
• Lean Management
• Android Development
• Problem Solving
• Deep Learning