Saqib Azim

Experience

• 

  Mood Indigo IIT Bombay

  Aug 2016 - Dec 2016

  Web Coordinator

  Contributed to the development of Mood Indigo '16 website

• 

  Indian Institute of Technology, Bombay

  Apr 2017 - Sept 2017

  Undergraduate Student Researcher

  - Worked under the supervision of Prof. Subhasis Chaudhuri at VIP Lab, IIT Bombay focusing on Zero-Shot Learning for Object Recognition.- To address this challenge, I proposed and developed a semi-supervised VGG16 autoencoder model, which effectively learned visual-semantic mapping utilizing Word2Vec features. - As a result of my contributions, I successfully enhanced the performance of ZSL's unseen class on the Animals with Attributes dataset from 58.7% to an impressive 65.3%.

• 

  Unmesh Mashruwala Innovation Cell

  Aug 2017 - Apr 2018

  Technical Team Member

  Worked as part of the Autonomous Driving Car team (SeDriCa) under Innovation Cell, IIT Bombay. Focused on solving problems related to the detection of roads, lanes, obstacles under varying lighting conditions, especially targeted at Indian driving conditions.

• 

  Samsung India

  May 2018 - Jul 2018

  Summer Intern

  - As part of the text recognition team at the Advanced Technology Lab, I played a key role in developing a 3D handwritten text recognition system that estimated wrist and hand movements using Samsung smartwatch IMU sensors. - One of the major challenges was modeling raw sensor noise, causing significant drifts in the generated character patterns. To mitigate this issue, I implemented adaptive frequency filters to preprocess the raw signals and improve the signal-to-noise ratio. - Designed the framework for data collection, and trained a SVM and LSTM model to learn the relation between hand movements and character patterns. - Achieved an impressive 93% text recognition accuracy on unseen data. Show less

• 

  Indian Institute of Technology, Bombay

  Jan 2019 - Jul 2019

  - Designed a novel control algorithm to drive a multi-agent system to a target destination.- Utilized global iterative solvers to estimate optimal paths in constrained conditions.- Learned to accurately predict pursuer-evader trajectories using an attention-LSTM model. EE 210 - Signals and SystemsAssisted Prof. Jayakrishnan Nair in the evaluation of assignments, exams, conducting office-hours, etc.

  • Undergraduate Student Researcher

    Aug 2018 - Jul 2019

  • Teaching Assistant

    Jan 2019 - Apr 2019
• 

  Hitachi

  Oct 2019 - Sept 2021

  Assistant Researcher

  Worked in the Intelligent Vision Research Group on a diverse range of AI and computer vision problems such as Visual Localization and Mapping, SLAM, Camera Tracking, Navigation, Object detection and segmentation, Hazardous Activity detection, etc.

• 

  UC San Diego

  Jan 2022 - Aug 2023

  DSC 140A - Probabilistic Modeling and Machine Learning CSE 166 - Image Processing ECE 225A - Probability and Statistics for Data Science Worked under the supervision of Prof. Lily Weng on Adversarial Robustness Analysis of Deep Models. Conducted an empirical analysis of the CLIP model's resilience to adversarial perturbationsand devised an attack mechanism to generate adversarial examples. Trained robust classifier with strong provable guarantees against adversarial attacks. ECE 109 - Probability and Statistics ECE 101 - Linear Signals and Systems

  • Graduate Student Researcher

    Oct 2022 - Aug 2023

  • Graduate Teaching Assistant

    Apr 2023 - Jun 2023

  • Graduate Teaching Assistant

    Jan 2023 - Mar 2023

  • Graduate Teaching Assistant

    Sept 2022 - Dec 2022

  • Graduate Student Researcher

    Apr 2022 - Aug 2022

  • Graduate Teaching Assistant

    Mar 2022 - Jun 2022

  • Graduate Teaching Assistant

    Jan 2022 - Mar 2022
• 

  San Diego Supercomputer Center

  Nov 2023 - now

  AI Research Engineer