Services
Video meeting . 15 mins
Priority DM . 2 days reply
Video meeting . 30 mins
Video meeting . 30 mins
Video meeting . 30 mins
Priority DM . 2 days reply
Priority DM . 2 days reply
Video meeting . 30 mins
Video meeting . 30 mins
Video meeting . 60 mins
About me
🚀 Lead Data Scientist (VP) @JPMorgan Chase | LLM Agents 💡🔬🔍🔧🛠️ | Generative AI 🧠 | LLM Enthusiast | 3AI Emerging AI Leader Award 🏆 | NLP Researcher | Data Centric AI | Open Source Contributor
A problem solver at heart, Ankur Kumar strives to build creative and effective solutions through cutting-edge AI techniques. Innovative AI and Machine Learning Expert with ~8 years in AI, specializing in LLMs and GenAI. A 3AI Emerging AI Leader with a strong problem-solving focus, successfully developed and deployed three generative AI projects into full production. Key achievements include ranking 6th in the Stanford Question & Answering Dataset 2.0 challenge and pioneering the first-ever algorithm to complete all 20 Facebook QA bAbi tasks with 100% accuracy.
Working On:
• Making NLP Products based on Question/Answer, Phrase Extraction, Summarization, and Classification.
• Research on Natural Language Representation.
• Productionizing Models using AWS ECS, AWS Sagemaker, AWS Lambda, AWS Step Function.
• Making D3.js based NLP Product visualization.
Worked On:
• Natural Language Understanding system and Our Team Solved Facebook BABI Task(NLP benchmarking) with 100% accuracy
• Made Image Classification and Object Detection Models
• Used ResNet Model for Image Classification and tried RCNN and its variants, SSD & YOLO Models for Object Detection.
Data Science Skills:
Natural Language Processing, Deep Learning, Machine Learning, Time Series Forecasting, Data Visualization, Statistics
Programming Language Skills:
Python, R, Java, C
Big Data Skills:
Good Knowledge and Experience of Hadoop, Pig, Hive, Sqoop, Flume, ELK Stack
Global NLP/NLU Benchmarking Standings:
a. Global AI Benchmark SQuAD 2.0 (The Stanford Question Answering Dataset)- ABG DnA 2020
• Created a Q/A model which ranked 6th (Feb-March’20) worldwide in the Global AI Performance Benchmark on Natural Language Understanding beating algorithms created by technology behemoths like Google Brain, Facebook
AI Research, and Microsoft Research Asia.
b. Facebook QA bAbi Task by Facebook AI Research (FAIR)- Dataval Analytics 2017
• First in the world to complete all 20 Facebook QA bAbi tasks with 100% accuracy.
My Top Performance on Machine Learning Competition:
• Secured 1st Prize award in NLP Hackathon GitHub Bug Prediction Challenge.
• Won Kaggle Weekly Kernel Award on August 31, 2018
• Kaggle Mercari Price Suggestion Challenge 2018 - Top 7%
• Analytics Vidhya Knocktober 2016 - Top 3%
• Analytics Vidhya The Ultimate Student Hunt 2016 - Top 4%