Shubham Navin Pathak

Software Engineer

Hyderabad, Telangana, India4 yrs 7 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

Expert in optimizing deep learning models for hardware.
Strong foundation in AI Infrastructure and Data Science.
Proficient in multiple programming languages and frameworks.

Stackforce AI infers this person is a Deep Learning and AI Infrastructure specialist with a focus on performance optimization.

Contact

Skills

Core Skills

Deep LearningAi InfrastructureData ScienceArtificial IntelligenceWeb DevelopmentMachine Learning

Other Skills

Open neural network exchange (Onnx)NumPyC++CPPPython (Programming Language)JavaScriptPythonDjangoKerasTensorFlowProfilingPresentation SkillsInterpersonal CommunicationPerformance BenchmarkingParallel Programming

About

Currently working in AI Engine team at AMD as an AI Infrastructure Engineer (Embedded applications), focused on optimizing deep learning models for efficient inference on next-generation hardware. My interests span Statistics, Deep Learning, Parallel Computing, Computer Architecture and Operating Systems, with a strong focus on performance optimization driven by numerical methods, programming and hardware-aware techniques to drive scalable, high-performance computing solutions. I’ve also dabbled in Data Science, Business Intelligence and Web Development, and enjoyed the process along the way :)

Experience

4 yrs 7 mos

Total Experience

1 yr 6 mos

Average Tenure

1 yr 11 mos

Current Experience

Amd

2 roles

Software Development Engineer 2

Jul 2024 – Present · 1 yr 11 mos · On-site

Worked on the bringup of performant models ( CNNs, Transformer-based architectures, and Generative models (Inception, ResNet variants, YOLO, ADAS etc)) across INT8, BF16, BFP16, and Micro-scaled instruction formats for benchmarking and / or as a part of compiler infrastructure, for Client and Embedded applications.
Key responsibilities :
Performing end-to-end ONNX model manipulation, including graph-level edits such as layer fusion and custom operator integration; data extraction and pre / post-processing transformations.
Creating python bit exact models.
Contribute to the frontend ML compiler components like the compute scheduler and quantizer.
⁠Working on tiling, data movement strategies and memory access patterns across hierarchical memory designs for AI workloads on single AIE core.
Developing optimised kernels and single core testbenches for MatMul, non-linear, element-wise and interpolation operations, for Inference library as well as for custom ops for Vision models to be run on multi-core systems; improving performance and memory use across the board.
Design test, profiling, and debugging infrastructure to aid in the bring-up, tuning, and debugging of DNN inference accelerators.

Open neural network exchange (Onnx)NumPyDeep LearningAI Infrastructure

Co-op / Intern

Jan 2024 – Jun 2024 · 5 mos · On-site

Open neural network exchange (Onnx)C++

Birla institute of technology and science, pilani

Teaching Assistant, Student Researcher

Aug 2022 – Dec 2023 · 1 yr 4 mos · India · On-site

Worked as a Teaching Assistant 👨‍🏫 across semesters for Foundations of Data Science (CS F320), Artificial Intelligence (CSF407), and Operating Systems (CSF372), helping students navigate labs and practicals, conducting vivas, and grading assignments and exams.
Juggled that with some research 👨‍🔬 🖥️
Worked on SARSA-based reinforcement learning prefetching model, helping systems anticipate what’s coming next instead of waiting around.
Also worked on optimizing load balancing and TSV placement in 3D NoC architectures, fine-tuning the hidden data highways inside processors.

CPPPython (Programming Language)Data ScienceArtificial Intelligence