Experience

Analysing biases in NLP models and data in Indian Context [AACL 2022]

Mitigating India-specific biases in encoder models using counterfactual data [In-progress]

Developed a pipeline to generate interpretable evaluation suites at scale for multiple tasks and languages. [Findings of AACL 2022 2022]

Worked on evaluation & improvement of NLP systems using human-in-loop techniques. [HumEval at EACL’21]

Wrote my bachelor thesis on the limitations and capabilities of large language models. [ICON 2021]

Worked on multimodal fake news detection integrating language, vision, metadata, source credibility, and explanations.

Achieved state-of-the-art performance on the LIAR-Plus and FakeNewsNet datasets. [ICDMAI 2021]

Automated an accounting process reducing the service time from 2-3 hours to less than one minute.

Built UI using TypeScript with React & MobX & backend services using Java-11 with Spring & MyBatis.

Developed an application for Detection of Changes in Land Usage using Remote Sensing Imagery & ML.

Integrated web-services to obtain satellite data from Sentinel Hub (a cloud-based GIS Platform).

Used OpenCV & scikit-learnto detect changes using Structural Similarity Index and k-NN methods.

Built a mobile application for logging work hours, leave registrations and schedule sharing for employees.