Description :

We are a Global Semiconductor Company, setting R&D team in Bangalore, India.

Position name : AI Software Stack Engineer

Work Location : Bangalore, India

We are building a next-generation software stack designed for high-performance execution on custom hardware. Our mission is to deliver industry-leading low-latency systems by developing an optimized, modular, and deeply integrated platform from workload ingestion to hardware-level execution.

We are looking for strong engineers who will serve as real user of our AI software stack. You will exercise the full stack end-to-end from ML frameworks and serving platforms through kernel authoring and compilation validating functionality, performance, and usability from the perspective of ML engineers and infrastructure teams. Your feedback will directly shape the quality and developer experience of the platform. This is an engineering position.

Key Responsibilities :

- Validate the AI software stack end-to-end by developing and running representative ML workloads, acting as a proxy for real users.

- Write from scratch, port, and optimize AI workloads, identifying gaps, performance bottlenecks, and usability issues.

- Develop and maintain a comprehensive suite of validation tests spanning model serving, kernel compilation, and framework integration layers.

- Reproduce, triage, and characterize issues across the stack from Python-level framework behavior down to compiled kernel correctness and performance.

- Benchmark and profile workloads to track performance regressions and validate optimization improvements across stack releases.

- Collaborate closely with compiler, runtime, and framework teams to provide actionable feedback and drive resolution of identified issues.

- Design and implement testing frameworks, platforms, and automation infrastructure to enable continuous and scalable validation across the stack.

Qualifications:

- BSc or higher in Computer Science, Electrical Engineering, or a related field.

- Hands-on experience with ML frameworks such as PyTorch, including model authoring and debugging.

- Familiarity with model serving platforms.

- Experience writing or modifying GPU kernels using Triton, CUDA, or similar kernel authoring tools.

- Strong Python skills and comfort working across multiple layers of a complex software stack.

- Systematic debugging mindset with the ability to isolate issues across framework, compiler, and runtime boundaries.

Strong Advantage :

- Experience enabling new model architectures or workloads on AI accelerator platforms.

- Hands-on experience with performance profiling and benchmarking tools for ML workloads.

- Understanding of compiler-generated code behavior and ability to read and reason about IR-level representations.

- Experience with CI/CD pipelines and automated test infrastructure for ML systems.

- Exposure to GPU or custom accelerators ecosystems.

- Familiarity with container-based deployment and orchestration for ML serving.