Omkar Joglekar

• Researching Foundation Models with a focus on Reasoning and Representation Learning.

• Worked on developing a Gradient-free training protocol for Binary and Ternary Weight models that enables training and fine tuning of large models on edge devices.
• Worked on developing a Foundation Model for Automated Robotic Assembly using text-prompt
• Developed a Text-Promptable solution for Instance Segmentation of Deformable Linear Objects such as wires, pipes, ropes and cables.

• Mentored two students from the Tel Aviv Youth University's Alpha program
• Co-developed a multi-threaded simulation of a Heterogeneous Robot Swarm
• Designed a Heuristic-based scheduling pipeline to perform exploration and mining on extraterrestrial planets with this swarm under full, partial and zero-communication settings

• Design and teach a course on "Intro to Machine Learning and AI" for middle school students
• Created an interactive GUI for the students to experience hands-on model training and inference.
• Created lecture slides, collected demos and videos, interactive theory explanations to get students excited about the field

• Designing experiments to prove the benefits of modular Control Systems over End-to-End pipelines, in terms of state estimation, feedback and fault detection in autonomous systems

• Built a pipeline to control and simulate the effect of Memory Approximation in RISC-V processor RAMs, within the gem5 processor simulation environment.

• Build a CNN-based framework to speed up the detection of Urinary Tract Infections (UTIs).

Graduated with honors. Specialized in Machine Learning and Data Science.

GPA: 93.1%

Graduated Valedictorian with Summa cum Laude.

GPA: 95.37%