Experience

Stealth Mode Startup

• Public Health

• Artificial Intelligence Engineer

  • Jun 2023 - Present

  CV, NLP, LLMs, VLMs. Using the state of the art to build new products on the edge. CV, NLP, LLMs, VLMs. Using the state of the art to build new products on the edge.



Purdue University

• United States
• Higher Education
• 700 & Above Employee

• Undergraduate Research Assistant @ CoRAL Laboratory

  • Jan 2023 - Present

  Working on interacting with robots through natural language.

• Undergraduate Research Assistant @ AKRaNLU Laboratory

  • Jan 2023 - Aug 2023

  Task-oriented multimodal conversational agents research at Purdue's AKRaNLU Laboratory. Selected to participate in the Alexa Prize TaskBot Challenge 2. I was an equal contributor to a research proposal that won $250,000 from Amazon Alexa.



Amazon

• Bangladesh
• Advertising Services
• 1 - 100 Employee

• SDE Intern

  • Aug 2022 - Nov 2022

  • Health Storefront and Tech Applied Science team, AI R&D • Built an NLP-driven & customer-facing recommendation engine for health guidance • Presented to Amazon Health leadership, helped plan product launch and future product goals • Health Storefront and Tech Applied Science team, AI R&D • Built an NLP-driven & customer-facing recommendation engine for health guidance • Presented to Amazon Health leadership, helped plan product launch and future product goals



Shield AI

• United States
• Software Development
• 400 - 500 Employee

• Artificial Intelligence Intern

  • May 2022 - Aug 2022

  • Developed a malleable end-to-end supervised ML pipeline with custom models, losses, data pipelines, & metrics • Built the foundation for the team to explore various offline deep reinforcement learning algorithms, including upside down reinforcement learning and behavior cloning • Ran experiments for knowledge distillation of existing AI pilots to build baseline pilot models • Developed a malleable end-to-end supervised ML pipeline with custom models, losses, data pipelines, & metrics • Built the foundation for the team to explore various offline deep reinforcement learning algorithms, including upside down reinforcement learning and behavior cloning • Ran experiments for knowledge distillation of existing AI pilots to build baseline pilot models

• Tensorflow Model Developer | GoogleML x Purdue University

  • Jan 2021 - May 2022

  • Leading the development of cutting-edge deep learning computer vision models and complementary tutorials for TensorFlow 2.x’s Model Garden to be used by the greater machine learning community • Increased the average precision of our YOLOv4-tiny model from 16% to 21.2% through hyperparameter optimization (https://github.com/tensorflow/models/pull/10605) • Created a TF-Record generator from the Pix3D dataset for the development of a 3D mesh predictor model (https://github.com/tensorflow/models/pull/10373) • Found a bug within TensorFlow where evaluating popular models on TPUs causes out-of-memory crashes (https://github.com/tensorflow/models/issues/10528) Show less



Pacific Northwest National Laboratory

• United States
• Research Services
• 700 & Above Employee

• Technical Intern

  • May 2021 - Aug 2021

  • Implemented linear-algebra based graph algorithms in NumPy, TensorFlow, and PyTorch • Performed an analysis of the ease of use and performance of different frameworks to gain insight into the development of a domain-specific programming language • Presented my analysis to peers and mentors during an intern research symposium • Implemented linear-algebra based graph algorithms in NumPy, TensorFlow, and PyTorch • Performed an analysis of the ease of use and performance of different frameworks to gain insight into the development of a domain-specific programming language • Presented my analysis to peers and mentors during an intern research symposium



Brookhaven National Laboratory

• United States
• Research Services
• 700 & Above Employee

• High School Research Intern

  • Jul 2020 - Aug 2020

  Under the mentorship of Dr. Huub Van Dam, my lab group and I analyzed the structure and potential mechanisms for inhibition of SARS-CoV-2 accessory protein ORF7a using computational docking methods and high performance computing resources. • Used DOCK6 and Autodock4 to screen potential ligands for inhibition of ORF7a's active sites. • Utilized GROMACS to investigate how the best ligand-ORF7a complexes interacted over time. • Performed structural analysis on ORF7a using PyMol and UCSF Chimera to observe which moieties and functional groups are important for future drug design. Show less