CV

Education

• M.S. Computer Science, Courant Institute of Mathematical Sciences, New York University, 2018
• B.Tech. Electronics & Comm. Engineering, Delhi Technological University, 2015
• Continuing Education, Stanford University, 2024
  • Deep Generative Models course offered by Stanford School of Engineering

Work Experience

• Senior Applied Scientist / ML Engineer – Instacart (Sept. 2025 – Present)
  • Logistics ML / AI team, New York, NY
  • Deep learning models for accurate delivery ETA prediction for the Instacart grocery delivery service.
• Applied Scientist II / Applied Scientist – Microsoft (June 2020 – July 2025)
  • Bing Maps AI team, Redmond, WA
  • Geospatial representation learning using vision-language foundation models to derive multimodal embeddings from aerial/street-view imagery and points-of-interest data. Object detection in street-view imagery. Semantic segmentation for error detection in digital surface models. Statistical modeling for traffic prediction. Built an LLM-powered agentic system for urban navigation. Published research at ICLR and ICML workshops and won Best Technical Hack at the Microsoft Global Hackathon 2021.
• Software Engineer – Amazon (Jan. 2018 – June 2020)
  • Prime Video Compliance & Classification / Operations Finance Tech, Seattle, WA
  • Scene-boundary detection and video copyright infringement classification using video and audio. APIs, storage, and metrics monitoring solutions. Published a US patent on temporal localization in long-form video.
• Software Engineer Intern – Amazon (May 2017 – Aug. 2017)
  • Operations Finance Tech, Seattle, WA
  • Shipped a statistical analysis tool to production; received return offer as full-time SDE.

Publications

Subimage Overlap Prediction: Task-Aligned Self-Supervised Pretraining For Semantic Segmentation In Remote Sensing Imagery

Lakshay Sharma, Alex Marin. (2026). "Subimage Overlap Prediction: Task-Aligned Self-Supervised Pretraining For Semantic Segmentation In Remote Sensing Imagery." WACV 2026 Workshop: CV4EO.

Landsat-Bench: Datasets and Benchmarks for Landsat Foundation Models

Isaac Corley, Lakshay Sharma, Ruth Crasto. (2025). "Landsat-Bench: Datasets and Benchmarks for Landsat Foundation Models." ICML 2025 Workshop: TerraBytes.

SuoiAI: Building a Dataset for Aquatic Invertebrates in Vietnam

Tue Vo, Lakshay Sharma, Tuan Dinh, Khuong Dinh, Trang Nguyen, Trung Phan, Minh Do, Duong Vu. (2025). "SuoiAI: Building a Dataset for Aquatic Invertebrates in Vietnam." ICLR 2025 Workshop: Tackling Climate Change with Machine Learning.

Neural Image Captioning

Elaina Tan*, Lakshay Sharma*. (2019). "Neural Image Captioning." arXiv preprint. (*Equal contribution)

Natural Language Understanding with the Quora Question Pairs Dataset

Lakshay Sharma*, Laura Graesser*, Nikita Nangia*, Utku Evci*. (2019). "Natural Language Understanding with the Quora Question Pairs Dataset." arXiv preprint. (*Equal contribution)

Sentiment Classification using Images and Label Embeddings

Laura Graesser, Abhinav Gupta, Lakshay Sharma, Evelina Bakhturina. (2017). "Sentiment Classification using Images and Label Embeddings." arXiv preprint.

Patents

• Temporal Localization of Mature Content in Long-Form Videos Using Only Video-Level Labels. US Patent 11829413. Patent
• Generating Improved Traffic Speed Data for Road Segments in a Geographical Area Using Traffic Speed Prediction Neural Networks. Patent in filed/pending status (United States).

Awards

• Winner of Best Technical Hack @ Microsoft Global Hackathon 2021 (for Subimage Overlap Prediction work)
• 1st place winner @ AWS DeepLens Hackathon for trash classification computer vision model; featured on AWS Blog and AWS YouTube
• Placed 13/96 in iMet2020 dataset classification challenge at CVPR 2020 Workshop on Fine-Grained Visual Categorization
• Toyota Motor Corporation Scholarship, awarded by International House NYC and Toyota

Service

• Peer reviewer / program committee member for multiple workshops/conferences including AAAI 2025, AAAI 2026, WACV 2025