SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Tenstorrent is building cutting-edge AI technology that unifies innovations in software models, compilers, platforms, networking, and semiconductors. The company has developed a high-performance RISC-V CPU from scratch and is growing its team of technologists passionate about solving hard problems in AI infrastructure.
Join Tenstorrent's AI IP software team to help customers bring AI workloads to products built on Tenstorrent technology. Beyond building internal hardware platforms, Tenstorrent offers AI acceleration IP that customers integrate into their own SoCs. You will work directly with customers to help bring models into Tenstorrent workflows, answer technical questions, and solve integration challenges. This is a broad, hands-on software role spanning models, compiler and runtime workflows, and AI IP.
You will learn how customers integrate Tenstorrent AI IP into their products and bring real-world models through the software stack. You'll gain experience with compiler and tt-nn workflows as they evolve, and understand how to validate model integrations. You'll see firsthand how compiler, runtime, software, IP, and hardware teams collaborate to improve model compatibility and customer outcomes, while solving practical customer challenges that contribute to high-performance AI technology for a wide range of products.
The role is hybrid, based out of Belgrade, Serbia. Tenstorrent welcomes candidates at various experience levels; during the interview process, candidates will be assessed for the appropriate level, and offers will align with that assessment.
**Requirements:**
- Strong C++ skills and experience developing, debugging, or maintaining production-quality software
- Hands-on experience with ONNX or similar model formats and workflows, such as model conversion, graph inspection or transformation, operator support, or runtime integration
- Experience bringing models into a software runtime or hardware platform and validating that integrations produce correct results
- Ability to work across compiler and software layers, investigate model compatibility, and help address customer-specific requirements
- Working understanding of machine learning fundamentals and common model architectures, especially computer vision models and large language models (expertise in every architecture not required)
- Strong problem-solving skills and ability to turn customer questions and ambiguous requirements into clear investigations and workable solutions
- Collaborative communication skills and experience working directly with customers and cross-functional engineering teams
- Plus: experience with compilers, MLIR, graph optimization, runtimes, code generation, customer-facing engineering, or technical support
