# Ishan Kumar > SWE at Google | MS CS at Columbia University | ex Applied Scientist at Microsoft | IIT Roorkee ’22 Location: New York, New York, United States Profile: https://flows.cv/ishankumar I am a second-year Master in Computer Science student at Columbia University focussing on ML. I read that over 90% of ML models never see the light of production and decided I want to change that. I focussed my courses at Columbia to learn more about LLMs as well as specialise in deployments (like MLOps) and inference-time optimisations for ML. I love thinking about open problems and building products. That's why working in applied research, where we leverage state-of-the-art models to solve actual user needs, really excites me. I am keen to learn more through internships in the industry Previously, I was a Applied Scientist at Bing, Microsoft for 2 years. I also worked as a researcher in the NLP lab at ETH Zurich. I graduated from the Indian Institute of Technology Roorkee in 2022. I am open to full time roles in Machine Learning, Research Engineer, Applied Scientist, Data Scientist and alike. ## Work Experience ### Software Engineer (AI/ML) @ Google Jan 2026 – Present | Mountain View, CA ### Research Assistant @ Columbia Engineering Jan 2024 – Present | New York, New York, United States Improving sampling for Diffusion LLMs like LLaDA with Prof Goldblum's group. - Looking at past predictions for positions - Improving KV Caching for diffusion LLMs. ### Applied Science Summer Assosciate @ JPMorganChase Jan 2025 – Jan 2025 | Seattle, Washington, United States Developed novel LLM Eval methods for RAG pipelines at the Machine Learning Center of Excellence. ### Data and Applied Scientist @ Microsoft Jan 2022 – Jan 2024 | Hyderabad, Telangana, India Worked as an Applied Scientist for 2 yrs at the Search Technology Center India. • Trained and optimized XLM-R models, cutting L1 latency from 200ms+ to 12ms at 20B+ webpage scale, generating 1M+ new events. • Applied optimizations including ONNX conversion, model pruning, and Blingfire tokenization, deploying models to serve live production traffic. • Implemented end‐to‐end data refresh pipeline, resulting in a 1.5M DAU gain • Reduced hallucinations by 15.7% with SegRank, a Novel LLM Grounding method • Won the Award of Excellence for Innovation in Microsoft, IDC for my work on the events datapipeline. • Honourable Mention at the Executive Challenge in the Microsoft Global Hackathon 2024 ### Research Assistant @ ETH Zürich Jan 2023 – Jan 2023 | Zurich, Switzerland • Formulated a novel metric CausalCite: a causal formulation of paper citations • Conducted highly parallelized experiments on a Citations graph with 2B+ edges and 200M+ nodes • Introduced innovative evaluation metrics with 30.14% better correlation with test-of-time than citations Published at ACL 2024, Bangkok ### Machine Learning Intern @ Zomato Jan 2021 – Jan 2021 | Gurugram, Haryana, India Machine Learning for new user Fraud Detection leveraging app events data. ### Open Source Contributor @ PyTorch Ignite Jan 2021 – Jan 2021 Ignite is a high-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently. Some of my notable contributions to the library:- • Added state attribute logging to all supported loggers. • Added batch support and micro averaging for Bleu and Rouge Metric. • Added tutorial for implementing seq2seq models using Ignite. https://pytorch-ignite.ai/tutorials/intermediate/02-machine_translation_using_pytorch_ignite/ Contributors list: https://github.com/pytorch/ignite/graphs/contributors ### Machine Learning Intern @ Deloitte India (Offices of the US) Jan 2021 – Jan 2021 AI Center of Excellence in Deloitte. • Built a model to predict the Probability of Success of a pursuit for an internal CRM tool at Deloitte. • Worked with the platform team on the internal ML platform. ### Research Student @ Video Analytics Lab Jan 2020 – Jan 2020 | Indian Institue of Science Bengaluru, Karnataka Working on Hand Pose recovery from single 2D RGB image, doing so in a Self Supervised manner. This will enable us to use the large number of unlabelled data available on the internet as well as make the model more robust since its trained on real world outdoors data also. ### Research Student @ CNERG - Complex Networks Research Group, IIT Kharagpur Jan 2020 – Jan 2020 | Indian Institute of Technology Kharagpur ## Education ### Master of Science - MS in Computer Science Columbia University ### Bachelor of Technology - BTech Indian Institute of Technology, Roorkee ### Leadership Program in Consulting ShARE - Growing a new generation of leaders ### Secondary School Springdales School, India ### Middle School NIST International School ### Amazon ML Summer School ## Contact & Social - LinkedIn: https://linkedin.com/in/ishan-k216 --- Source: https://flows.cv/ishankumar JSON Resume: https://flows.cv/ishankumar/resume.json Last updated: 2026-04-13