Experience
-
Associate ML Scientist IIJan 2025 – Sep 2025Wadhwani Institute for Artificial Intelligence, DelhiComputer-vision models for crop monitoring (joint positional embeddings, MAE-style pretraining); fine-tuned open-source VLMs for agriculture; led pest-monitoring projects and built experimental frameworks for low-resource, noisy-data settings.
-
Associate ML Scientist IJul 2023 – Jan 2025Wadhwani Institute for Artificial Intelligence, DelhiBuilt standardized training/eval and deployment pipelines (~80% faster); improved robustness with dataset cartography and CAM explainability; developed an auditable crop-yield estimation system; shipped CottonAce with synthetic image generation for national deployment in the Ministry of Agriculture's National Pest Surveillance System.
-
ML InternFeb 2023 – Jul 2023Wadhwani Institute for Artificial Intelligence, DelhiDeveloped and evaluated novel model-distillation techniques for object detection and counting; improved cascade models (including OOD detectors) for edge devices; contributed to the pest-infestation analysis pipeline.
-
Research InternAug 2022 – Feb 2023IIT Delhi — Prof. Chetan Arora, Prof. M. BalakrishnanMedical AI for gall bladder cancer detection from ultrasonography; Vision Transformer explainability, weakly supervised object detection, and multiple-instance learning; model optimization for the MAVI assistive-vision project.
-
Summer Research FellowJun 2022 – Aug 2022IIT DelhiOptimized deep-learning models for scene-text recognition on edge devices; upgraded models to OpenVINO 2022 / PyCam 2.0; converted and tested VisionLAN (ViT + language model) to ONNX and OpenVINO IR.
For education, awards, and technical skills, see my CV.