Meta Reality Labs, Burlingame, CA
Meta Reality Labs, Burlingame, CA
Research Engineer (June 2024 - Current)
Embedded Files
I work on deploying cutting-edge 3D rendering research for digital avatars in AR/VR products.
Carnegie Mellon University, US
Carnegie Mellon University, US
Research Assitant advised by Prof. Katia Sycara (Sept 2022 - Current)
Developing interpretable ML models inspired by human brain sub-architectures for the task of real-time decision-making on sequential online data; currently focusing on action recognition in live-stream videos
The model accumulates evidence on the fly and makes online predictions based on a learned threshold model
ETH Zurich, Switzerland
ETH Zurich, Switzerland
Research Intern with Prof. Mrinmaya Sachan (Feb 2022 - June 2022)
Created a framework in OpenAI Gym for teaching grammar to RL agents in an online policy learning context
Successfully trained an RL agent for grammar error correction using the A3C, demonstrating its ability to learn via self-exploration directly from incorrect sentences without prior supervision or grammar knowledge
Microsoft, India
Microsoft, India
Data & Applied Science Intern (May 2021 - Jul 2021)
Improved the quality of related search recommendations for Microsoft Bing as part of the Query Reformulation problem
Implemented a progressive SimRank algorithm for bipartite graphs to iteratively learn all query pair similarity scores
Improved the recall and quality on a Query-URL dataset (+4M nodes) by +21% DCG score over the baseline
Michigan State University, US
Michigan State University, US
Research Intern with Prof. Kalyanmoy Deb (May 2020 - July 2020)
Developed a robust ensemble-based crossover operator selection algorithm -EnXEA, for classical genetic algorithms
Updated the operator selection probabilities adaptively, using expectation maximization and temporal diff. learning
Achieved the state-of-the-art with minimal computation overhead & best IGD scores for 17 optimization problems
Developed a MATLAB Library for Non-dominated Sorting Genetic Algorithm (NSGA2 & 3) for open-source use
Nokia Networks, India
Nokia Networks, India
SDE Intern (April 2020 - May 2020)
Developed a deep CNN model for automated PDF data extraction using OCR to assign the correct metadata to pdf-content
Built a Git-style interface for file management on web-collaborative platforms- SharePoint & OneDrive using REST APIs
National Autonomous University of Mexico
National Autonomous University of Mexico
Research Intern (May 2019 - July 2019)
Designed an uncertainty weighted least-squares linear regression model to identify discordant outliers in multivariate datasets
Generated new critical values through Monte Carlo simulation to determine the probability of no-correlation criteria
Published results on “neonatal body measurement” data and showed significant improvement in correlation using our method of outlier detection with the uncertainty in the data decreasing by 25-40% compared to the traditional methods of statistics
Page updated
Google Sites
Report abuse