About Ziroh Labs

Ziroh Labs, a deep tech startup with a global presence spanning both the Bay Area and Bangalore, is at the forefront of innovation in the computing field. Ziroh Labs groundbreaking work involves the convergence of advanced mathematical concepts and sophisticated distributed system algorithms for the efficient execution of open-source foundational models on CPUs.  Through pioneering research and development, Ziroh Labs is playing a pivotal role in democratising artificial intelligence. This democratisation has the potential to break down the barriers that have limited access to AI technology, making it available to a broader range of individuals, businesses, and organisations.

Ziroh Labs, in collaboration with IIT Madras, recently launched the product Kompact AI and a Centre for Artificial Intelligence Research, where cutting-edge models are being developed to solve problems. You will contribute to Kompact AI development.

‍

Job Brief

We are seeking a highly motivated Dev0-level Systems Engineer with a strong foundation in modern C++ (C++17+), computer architecture, and low-level performance optimization. In this role, you’ll contribute to the development of high-performance software by leveraging processor intrinsics, implementing parallel algorithms using OpenMP and MPI, and optimizing critical compute routines like General Matrix Multiplication (GEMM). Ideal candidates will have hands-on experience with CPU instruction sets, a deep understanding of memory hierarchies and pipelines, and a passion for writing efficient, scalable, and hardware-aware code. This is an excellent opportunity to work at the intersection of systems programming, numerical computing, and AI infrastructure.

‍

Responsibilities

• Contribute to designing, developing, and maintaining the Java SDK for Kompact AI. 
• Create design documents, including architecture and data flow diagrams.
• Write well-defined and efficient code that meets coding standards and test cases. 
• Conduct code reviews to ensure adherence to coding best practices. 
• Debug and solve reported issues.
• Create a Java document outlining the functionality of the SDK.

‍

What Dev0 skills matter to us:

• Hands-on experience with processor intrinsics: Demonstrated ability to use specialized CPU instructions (intrinsics) to write highly optimized code, leveraging direct hardware capabilities for performance gains.
• Proficiency in parallel programming: Extensive experience in designing, implementing, and debugging concurrent and parallel applications using frameworks like OpenMP, MPI, to effectively utilize multi-core processors and distributed systems.
• Strong understanding of Computer Organization and Architecture: In-depth knowledge of CPU architecture, memory hierarchies (caches, RAM), instruction pipelines, and I/O mechanisms, enabling the development of efficient and hardware-aware software.
• Expertise in Modern C++ (C++17 and above): Advanced proficiency in modern C++ features, including but not limited to smart pointers, move semantics, lambdas, std::async, and std::thread, for writing robust, maintainable, and high-performance code.‍
• Understanding of General Matrix Multiplication (GEMM) algorithm: A solid theoretical and practical understanding of the General Matrix Multiplication (GEMM) algorithm is crucial. This includes knowledge of its computational complexity, various optimization techniques (e.g., blocking, cache-aware optimizations), and its critical role as a building block in designing and implementing AI model architectures. Experience with implementing and optimizing GEMM for different architectures is a strong plus.

‍

Interview Process:

• Meet and Greet 
• Technical Round (Virtual or F2F)
• HR

‍

Expected time for joining:

• Within 15-30 days post-selection.

‍

‍