Profile Picture

Aditi Tiwari

Ph.D. Student at UIUC

My name is Aditi, and I am a Ph.D. student at the University of Illinois Urbana-Champaign (UIUC), advised by Dr. Klara Nahrstedt and Dr. Heng Ji. My research is driven by a deep curiosity about exploring new ideas, applying existing concepts in completely novel ways, and bridging gaps between different fields. I thrive on solving complex problems and building AI systems that have real-world impact.

My academic journey has been shaped by exciting collaborations with Dr. Minjia Zhang, Dr. Derek Hoiem, Dr. Shenlong Wang, and Dr. Zhisheng Yan. Prior to my Ph.D., I completed my MSCS at UIUC in 2024 under Prof. Klara Nahrstedt, supported by the National Science Foundation. Over the years, I have gained research experience at Rivian, George Mason University (GMU), Coordinated Science Lab (UIUC), Indian Institute of Technology (IIT) Delhi, and Birla Institute of Technology And Science, Pilani (BITS Pilani).

My research interests lie in developing efficient and adaptive AI systems for multimodal learning, computer vision, and machine learning, including:

I am always open to exciting collaborations—if you’d like to discuss research, AI, or just have an interesting idea to explore together, feel free to reach out to me via email at aditit5 at illinois dot edu.

You can also find me on Google Scholar, LinkedIn, Medium, GitHub, and ResearchGate.

UIUC Research

Research Collaborator, UIUC

Jan 2025 - Present | Urbana, IL

  • Collaborating with Prof. Derek Hoiem & Prof. Shenlong Wang on feedforward rendering techniques for 3D scene understanding and multimodal vision-language tasks.
  • Exploring region-based slot attention for efficient 3D instance segmentation, improving semantic segmentation and object representation.
  • Investigating self-supervised learning for scene change detection and 3D instance retrieval, leveraging multi-view geometric consistency.
    • Rivian Automotive

    Research Intern (Computer Vision), Rivian Automotive

    May 2024 - Aug 2024 | Palo Alto, CA

  • Designed and coded custom Image Signal Processing (ISP) pipeline from scratch, converting raw sensor data into RGB/YUV formats.
  • Implemented core functions in C/C++ (histogram equalization, demosaicing, gamma correction) with 2-4s latency, surpassing company benchmarks.
  • Reduced manual input by 85% through automated tuning of image processing parameters.
    • Coordinated Science Lab

    Research Assistant (CV/ML), Coordinated Science Laboratory, UIUC

    Sep 2022 - May 2024 | Urbana, IL

  • Curated first action detection dataset for 360-degree firefighting videos, developed custom annotation scripts.
  • Developed ACT360, a quantized, low-latency action detection model surpassing YOWO with 4% mAP improvement (0.865) at 48ms/frame inference.
  • Optimized ACT360 for real-world deployment, reducing inference time to 33ms while maintaining 0.850 mAP.
  • Successfully deployed ACT360 at Illinois State Fire Department, making firefighter training analysis 95% faster.
    • George Mason University

    Research Collaborator, George Mason University

    Sep 2022 - Sep 2023 | Fairfax, VA

  • Conducted research under Dr. Zhisheng Yan on efficient real-time video analytics for resource-constrained environments.
  • Designed a low-latency action detection model for real-time surveillance applications, leveraging temporal filtering and multi-view learning techniques.
  • Evaluated model performance across IoT-enabled edge devices, achieving a 30% improvement in computational efficiency while maintaining high detection accuracy.
    • IIT Delhi

    Machine Learning Research Intern, IIT Delhi

    May 2021 - Sep 2021 | Delhi, India

  • Engineered UAV for autonomous object detection and route adaptation, achieving 88.21% accuracy (matching SOTA).
    • McAfee

    Software Engineer Intern, McAfee Inc.

    Aug 2021 - Dec 2021 | Bangalore, India

  • Optimized HAR file processing automation, reducing processing time by 58% using Python & Bash.
  • Built responsive customer query interface with AngularJS, JavaScript, cutting development time by 15%.
  • Automated SQL-based data extraction, saving 6 person-hours/week.
    • BITS Pilani

    Research Intern, Birla Institute of Technology and Science (BITS), Pilani

    Jan 2021 - June 2021 | Pilani, India

  • Worked under Dr. Vinay Chamola on the study of public key post-quantum cryptosystems.
  • Researched code-based, lattice-based, multivariate-based, supersingular elliptic curve isogeny, and hybrid cryptosystems to assess their feasibility in post-quantum security protocols.
  • Research findings contribute towards a pending research paper, focusing on quantum-safe cryptographic solutions for modern computing architectures.
    • MEDSupervision

    Full Stack Developer Intern, MEDSupervision (Startup)

    Nov 2020 - May 2021 | Delhi, India

  • Integrated real-time video calling using Twilio API, increasing user engagement by 37%.
  • Redesigned patient portal interface, improving usability for medical staff and patients.
  • Optimized database structure (~10,000 entries), achieving 50% faster data retrieval.
    • QikCircle

    Software Development Engineer Intern, QikCircle (Startup)

    Feb 2020 - May 2020 | Delhi, India

  • Developed front-end using HTML, CSS, JavaScript, boosting user engagement by 25%.
  • Improved backend functionalities with Python/Flask for seamless SQL-based data retrieval.
  • Designed user registration system, reducing account creation time by 40%.

    • Design inspired by Savya Khosla | Template on GitHub | Last updated: Feb 2025