HPC Architect - Bhubaneswar - Stealth AI Startup

Description

Job Description: HPC Architect

Stealth AI Startup is a well-funded company with an exceptional leadership team, dedicated to solving complex problems and pushing the boundaries of AI.

Responsibilities:

• Optimize and port scientific and AI/ML applications to harness maximum performance on HPC systems, including clusters, grids, and cloud platforms.
• Architect and evaluate diverse hardware architectures for performance optimization (e.g., GPUs, CPUs, and accelerators like Nvidia, AMD, Graphcore, and Cerebras).
• Spearhead benchmarking activities for evaluating HPC and AI applications using tools like MLPerf, LINPACK, and OSU benchmarks.
• Collaborate with research and engineering teams to understand application requirements and develop optimization strategies for parallel computing using CUDA, MPI, OpenMP, or OpenACC.
• Implement and manage HPC workloads using SLURM workload management systems.
• Lead the integration of HPC infrastructures with AI/ML pipelines for domains like computer vision, NLP, and LLMs using frameworks like TensorFlow, PyTorch, and OpenCV.
• Develop custom solutions for containerized deployments using tools like Docker, Kubernetes, Enroot, or Charliecloud.
• Mentor team members and provide technical guidance on HPC and AI solutions.

Qualifications:

• Educational Background: Bachelor's or Master's degree in Computer Science, Electronics, Telecommunications, or a related field.
• Programming Expertise: Proficiency in C/C++, Python, and shell scripting.
• Parallel Computing Skills: Expertise in CUDA, MPI, OpenMP, OpenACC, and optimization techniques.
• HPC Applications Knowledge: Experience with applications like GROMACS, LAMMPS, OpenFOAM, Quantum Espresso, or WRF.
• Benchmarking and Profiling Tools: Hands-on experience with HPL, HPCC, Intel VTune, Nvidia Nsight, or GProf.
• HPC Architecture Experience: Familiarity with CPUs (x86, ARM), GPUs (Nvidia, AMD), and accelerators.
• Ai/ML Frameworks: Practical experience with TensorFlow, PyTorch, Rapids, and Deepspeed for tasks like CNN, RNN, Transformers, NLP, and LLMs.
• DevOps Skills: Experience with containerization, cluster setup, and workload orchestration tools.
• Soft Skills: Excellent problem-solving, collaboration, and mentoring abilities.

Preferred Skills:

• Relevant Experience in HPC-AI research initiatives in Technology Companies.
• In-depth understanding of HPC infrastructure setup, including job scheduling systems like SLURM.
• Strong exposure to MLOps practices and deployment tools like HuggingFace, Gradio, or Streamlit.

What We Offer:

• Competitive salary and equity options.
• Opportunity to work with cutting-edge AI and HPC technologies.
• Collaborative, inclusive, and dynamic work culture.
• Work on impactful projects shaping the future of AI and supercomputing.