Senior Machine Learning Engineer — Autonomous Vehicles
NVIDIA Corporation, Santa Clara, CA
Jun 2021 - Present
- Improving perception DNN performance for autonomous vehicles using computer vision, machine learning, and synthetic data generation.
- Leveraging multi-camera perception systems, photorealistic datasets from Omniverse Replicator, and domain adaptation to enhance robustness across diverse environments.
- Led the development of ARSim, a synthetic augmentation framework that improved 3D localization accuracy by 10% and reduced prediction error by 20% across key perception tasks.
- Recently focusing on Generative AI–based synthetic data generation using diffusion models to expand long-tail scenario diversity.
Graduate Research Assistant
Georgia Institute of Technology, Atlanta, GA
Aug 2017 – Jun 2021
Conducted research on enabling edge intelligence in resource-constrained autonomous systems, with a focus on reinforcement learning applications. Developed energy-efficient ML algorithms (e.g., domain randomization), proposed novel hardware-aware architectures, and addressed scalability challenges in multitask federated reinforcement learning systems. Work focused on drone navigation and on-device learning.
Machine Learning System Intern
Samsung Semiconductor, San Jose, CA
May 2019 – August 2019
Worked in the Advanced Logic Lab on a novel STT-MRAM-based analog Processing-In-Memory (PIM) DNN accelerator. Developed an end-to-end simulation framework to optimize power-performance trade-offs for various DNN topologies. Achieved notable improvements in energy efficiency and latency over digital DNN accelerators.
Machine Learning Engineer — Autonomous Vehicles
Aivsol, Germany (Remote)
Jun 2017 – Aug 2017
Developed deep learning models for object detection and classification in self-driving car applications using TensorFlow. Focused on pedestrian, vehicle, and traffic sign recognition. Adapted state-of-the-art algorithms, trained models, and provided recommendations for optimal deployment strategies.
Publications#
[1] Aqeel Anwar, T. E. Choe, Z. Wang, S. Fidler, and M. Park, Augmented reality based simulated data (arsim) with multi-view consistency for AV perception networks, 2024. arXiv:2403.15370 [cs.CV]
[2] A. S. Lele, Y. Fang, Anwar, Aqeel, and A. Raychowdhury, “Bio-mimetic high-speed target localization with fused frame and event vision for edge application,” Frontiers in Neuroscience, vol. 16, 2022.
[3] Aqeel Anwar and A. Raychowdhury, Multi-task federated reinforcement learning with adversaries, 2021. arXiv:2103.06473 [cs.LG]
[4] A. K. Kosta, Anwar, Malik Aqeel, P. Panda, A. Raychowdhury, and K. Roy, “Rapid-RL: A reconfigurable architecture with preemptive-exits for efficient deep-reinforcement learning,” arXiv preprint arXiv:2109.08231, 2021.
[5] Z. Wan, Anwar, Aqeel, Y.-S. Hsiao, T. Jia, V. Janapa Reddi, and A. Raychowdhury, “Analyzing and improving fault tolerance of learning-based navigation system,” ACM/EDAC/IEEE DAC, 2021.
[6] A. Anwar and A. Raychowdhury, “Autonomous navigation via deep reinforcement learning for resource constraint edge nodes using transfer learning,” IEEE Access, vol. 8, pp. 26549–26560, 2020.
[7] A. Anwar, A. Raychowdhury, R. Hatcher, and T. Rakshit, “XbarOpt – enabling ultra-pipelined, novel STT MRAM based Processing-in-Memory DNN accelerator,” in IEEE AICAS, 2020, pp. 36–40.
[8] Aqeel Anwar and A. Raychowdhury, Masked face recognition for secure authentication, 2020. arXiv:2008.11104 [cs.CV]
[9] S. Zeng, Anwar, Aqeel, T. Doan, A. Raychowdhury, and J. Romberg, “A decentralized policy gradient approach to multi-task reinforcement learning,” arXiv preprint arXiv:2006.04338, 2020.
[10] I. Yoon, Anwar, Aqeel, T. Rakshit, and A. Raychowdhury, “Transfer and online reinforcement learning in STT-MRAM based embedded systems for autonomous drones,” in IEEE DATE, 2019, pp. 1489–1494.
[11] I. Yoon, Anwar, Malik Aqeel, R. V. Joshi, T. Rakshit, and A. Raychowdhury, “Hierarchical memory system with STT-MRAM and SRAM to support transfer and real-time reinforcement learning in autonomous drones,” IEEE Journal on Emerging and Selected Topics in Circuits and Systems, 2019.
[12] Anwar, Malik Aqeel and A. Raychowdhury, “Navren-RL: Learning to fly in real environment via end-to-end deep reinforcement learning using monocular images,” in IEEE M2VIP, 2018, pp. 1–6.
[13] Anwar, Malik Aqeel, A. B. Siddique, and M. Tahir, “Relative self-calibration of wireless acoustic sensor networks using dual positioning mobile beacon,” IEEE Systems Journal, vol. 12, no. 1, pp. 862–870, 2016.
Dissertation: Anwar, Malik Aqeel. "Enabling edge-intelligence in resource-constrained autonomous systems." Georgia Institute of Technology, 2021.
[P1] Anwar, Malik Aqeel, T. E. Choe, Z. Wang, S. Fidler, and M. Park, “Data augmentation for model training in autonomous systems and applications,” US Patent App. 18/592,025, Sep. 2024.
[P2] T. Rakshit, Anwar, Malik Aqeel, and R. Hatcher, “Batch size pipelined PIM accelerator for vision inference on multiple images,” US Patent 11,769,043, Sep. 2023.
[P3] T. Rakshit, Anwar, Aqeel, and R. Hatcher, “Ultra pipelined accelerator for machine learning inference,” US Patent App. 16/838,971, Apr. 2021.