Company: Tensordyne (formerly Recogni)
Role: Jr Software Engineer - ML Runtime - Rust
Location: Munich, Germany
Team: Low-Level Runtime Team

---

Company Overview

• Builds high-performance, low-power generative AI inference systems combining custom silicon, hardware, and software.
• Dual-headquartered in Sunnyvale, CA, and Munich, Germany, with distributed teams across North America and Europe.

---

Role Overview

• Works within the Low-Level Runtime team building components of a machine learning model execution framework.
• Focuses on developing a low-level assembler architecture optimized for proprietary hardware and porting model components for execution.

---

Responsibilities

• Maintain and develop low-level ML model execution frameworks in Rust.
• Maintain and enhance components and full models tailored for custom hardware.
• Optimize ML model execution to maximize performance and hardware utilization.
• Debug and productize models across virtual and physical hardware environments.

---

Requirements

Preferred Qualifications:

• Experience with the Rust programming language.
• General understanding of high-level ML model concepts (e.g., LLMs, attention architectures).
• Experience with low-level software development (e.g., embedded systems).
• Experience driving agents and building LLM engineering workflows.

Nice to Have:

• Experience with Google Cloud Platform (GCP) or equivalent cloud infrastructure.
• Experience with Python.
• Deep knowledge of modern LLM attention architectures (e.g., GDN, DSA, KDA).

---

Benefits & Workplace

• Comprehensive benefits package including competitive compensation, flexible spending options, and employee recognition programs.
• Equal Opportunity Employer (EOE).