---
title: 'Working Student - LLM Quantization & Inference Performance at Tensordyne'
canonical: 'https://feeny.ai/job/working-student-llm-quantization-inference-performance-tensordyne-munich-j7zc6yravggc'
type: 'job'
last_seen: '2026-11-06'
---

# Working Student - LLM Quantization & Inference Performance at Tensordyne

- **Company:** Tensordyne
- **Location:** Munich, Germany
- **Posted:** 2026-10-14
- **Last confirmed live:** 2026-11-06
- **Apply:** https://job-boards.greenhouse.io/tensordyne/jobs/8886436002

## Job description

## About Tensordyne (formerly Recogni)

AI is transforming our world. It can perform cognitive functions that previously only humans could do, such as perceiving interactions across different modalities and environments - with the ability to quickly learn and then solve complex problems. Tensordyne is an AI system solution company that builds very high-performance, low-power generative AI inference systems. Our mission, through the creation of custom silicon, hardware and software, is to enable multimodal Generative AI inference acceleration at scale, with safe, sustainable, high-performance systems for our hyperscaler and neocloud data center customers. We are at the leading edge of advancing the latest research and product improvements for generative Al inference solutions that will make Al even more advantageous for compelling new generative AI applications.Tensordyne is a well funded, fast-paced startup company with headquarters in both Sunnyvale, CA, and Munich, Germany. We also have many talented team members working remotely across North America and Europe. We take care of our people and their families with comprehensive benefits, competitive compensation, flexible spending options, and recognition programs, because building category-defining technology starts with a healthy, supported team. Come join us as we shape the future of multimodal generative artificial intelligence!

## Role summary

How much numerical precision does an LLM need, and where in the model architecture does it matter most? When does quantization improve runtime, and what does it cost in model quality?

We’re looking for a motivated working student to help turn these questions into experiments and experiments into maintainable code. You’ll explore optimization methods and build reusable evaluation tooling. You’ll work closely with our quantization, compiler, and runtime engineers, and your findings will directly influence how we optimize models for Tensordyne hardware.

## What you will do

- Model quality evaluation: Measure how low-precision computation affects performance on coding and reasoning benchmarks. Diagnose where quantization has the greatest impact.
- Quantization research: Implement and compare post-training quantization methods to improve inference efficiency while preserving task accuracy.
- Build reproducible tooling: Automate evaluations, track configurations and results, and make experiments accessible to engineers and AI agents.
- Measure runtime performance: Characterize latency, throughput, and memory usage to understand bottlenecks under different deployment scenarios.

A first project could involve exploring the quality–performance Pareto frontier of a Qwen or LlaMa model. You’ll experiment with 16-bit, 8-bit, and 4-bit quantization configurations to find the best runtime-quality tradeoff.

Core qualifications

- Current enrollment in Computer Science, Mathematics, Electrical Engineering, Machine Learning, or a related master’s degree program.
- Strong Python and PyTorch skills, including the ability to understand and modify model implementations.
- Effective use of AI coding tools and critical review of their output.
- An experimental mindset: you can formulate hypotheses, design comparisons, and draw careful conclusions from noisy results.
- Clear communication and the ability to present experimental findings and technical tradeoffs to the team.

Strong pluses: experience with quantization, LLM evaluation, performance profiling, or inference optimization.

This role is for you if you thrive on responsibility, learn quickly, and are naturally proactive.

## What we offer

Join a deep-tech team building the next generation of AI inference systems. Put your university knowledge into practice with leading open-source LLMs, deepen your understanding of quantization and inference optimization, and build skills in experimental design, software engineering, and AI coding tools. You’ll own a focused project, receive close mentorship from experienced engineers, and have room to explore ideas that shape our optimization methods.

Tensordyne's culture was built on the following values

- Put people first. We only succeed when our people succeed.
- Ethics and integrity always. Be open, honest, and respectful of everyone.
- Think Big. Be ambitious and have audacious goals.
- Aim for excellence. Quality and excellence count in everything we do.
- Own it and get it done. Results matter!
- Make each person better together than they would be as an individual.
- Embrace each other’s differences, and embrace that there will be differences.

Tensordyne is an equal opportunity employer. We believe diverse teams are better equipped to solve complex problems and build exceptional technology. All qualified applicants will receive consideration for employment without regard to age, color, gender identity or expression, marital status, national origin, disability, protected veteran status, race, religion, pregnancy, sexual orientation, or any other characteristic protected by applicable laws, regulations, and ordinances.

A note to recruitment agencies: Please do not contact Tensordyne employees or leaders regarding this role. We do not accept unsolicited agency resumes and are not responsible for fees associated with unsolicited submissions.

## About Tensordyne

## Company Overview
- **One-liner**: Tensordyne builds full-stack AI inference systems that use proprietary logarithmic math and custom 3nm silicon to dramatically reduce power, cost, and rack space for running large-scale generative AI models.
- **Entity Type**: Private (Series C funded)
- **Headquarters**: Sunnyvale, California, United States (co-headquarters also in Munich, Germany)
- **Founded**: 2018 (based on first funding round in 2018)
- **Founders**: Gilles Backhus (Co‑Founder, VP of AI & Product), RK Anand (Co‑founder & Chief Product Officer)

## Core Business
- **Primary industry**: AI hardware and infrastructure (computer hardware manufacturing, deep learning silicon, systems engineering)
- **Target customers**: Hyperscalers, Neo-Cloud data centers, and large enterprises running multimodal generative AI workloads
- **Mission**: “Innovating at the zeroth layer – rewrite the numbers so intelligence flows with a fraction of today’s energy and cost.”

## Products & Services
- **[Tensordyne Napier Inference System](https://www.tensordyne.ai/inference-system)**: An integrated rack‑scale AI inference platform combining custom logarithmic processors (3nm), HBM/SRAM memory, and a sub‑microsecond interconnect. Delivered as a TDN72 pod (72 chips) that scales into a full rack, supporting standard token serving and disaggregated inference. Fully air‑cooled and designed to fit into existing data center infrastructure.

## Market Standing
- **Valuation**: Not publicly disclosed
- **Key Metrics**: 
  - Annual Revenue: $4.5M (estimated, per LinkedIn)
  - Total Funding: $211.05M across 5 rounds
- **Notable Investors/Partners**: GreatPoint Ventures (led Series A), Celesta Capital (led Series B), Premji Invest (led Series C), HSBC Innovation Banking (debt financing)
- **Growth Signals**: Acquired Recogni GmbH; 89 employees (growing ~1% monthly); 12 active job postings (up 20% month-over-month, 140% quarter-over-quarter); talent drawn from top AI teams (Argo AI, Intel, NVIDIA, BMW, Bosch AI Center)

## Competitive Advantages
- **Logarithmic compute**: Replaces energy‑intensive multiplications with additions, cutting power consumption at the root.
- **Custom 3nm silicon** with built‑in interconnect logic, eliminating traditional bottlenecks.
- **Sub‑microsecond interconnect** (under 1,000 ns latency, 1 TB/s any‑to‑any bandwidth) enabling near‑linear scaling across 72 chips.
- **9× better space efficiency, 2× faster speed, and 10× cost savings** versus the Nvidia+Groq standard (company claim).
- **Fully air‑cooled** – no complex liquid cooling required, lowering deployment friction.

## Strategic Focus
Tensordyne is focused on making GenAI inference “profitable at speed and scale” by delivering a drop‑in system that dramatically reduces the total cost of ownership for hyperscale and cloud data centers. Current priorities include scaling production of the Napier system, expanding go‑to‑market with enterprise sales, and deepening the logarithmic quantization ecosystem to support any framework or model.

## Why Work Here
- **Culture**: “Build cool tech” – a band of engineers, mathematicians, and chip designers who thrive on solving hard, ambiguity‑rich problems. Values include one team, mission focus, innovate with passion, think like an owner, and act with integrity.
- **Remote/Hybrid**: Remote‑friendly; output over location. In‑person gatherings happen when “magic needs to happen.”
- **Engineering emphasis**: Deep technical challenges across math, silicon, systems, and software – with a fast‑moving, well‑funded environment. Team spans North America and Europe (offices in Sunnyvale, San Jose, Munich).
- **Hiring process**: Mutual discovery – a few calls, possibly a challenge, then fast decisions.

## Sources
1. [tensordyne.ai – About](https://www.tensordyne.ai/about)
2. [tensordyne.ai – Careers](https://www.tensordyne.ai/careers)
3. [tensordyne.ai – Inference System](https://www.tensordyne.ai/inference-system)
4. [linkedin.com – Tensordyne Company Page](https://www.linkedin.com/company/tensordyne)

## Other roles at Tensordyne

- [Working Student - LLM Inference Performance](https://feeny.ai/job/working-student-llm-inference-performance-tensordyne-munich-ep2f11b2p27s) — Munich, Germany
- [Senior Software Engineer — ML Inference & Numerics](https://feeny.ai/job/senior-software-engineer-ml-inference-numerics-tensordyne-munich-qrmjnkcnqpkp) — Munich, Germany
- [System Software / Embedded Engineer](https://feeny.ai/job/system-software-embedded-engineer-tensordyne-sunnyvale-1rwc6exvavsc) — Sunnyvale, CA
- [System Software/Embedded Engineer (Munich)](https://feeny.ai/job/system-software-embedded-engineer-munich-tensordyne-munich-ch5fc1wh9yvj) — Munich, Germany
- [Software Engineer – Cloud Services – Rust](https://feeny.ai/job/software-engineer-cloud-services-rust-tensordyne-munich-k2jksee9f5k3) — Munich, Germany
- [Ops/Developer Platform Engineer](https://feeny.ai/job/ops-developer-platform-engineer-tensordyne-munich-whbc1e2asynv) — Munich, Germany
- [Sr. Compiler Engineer](https://feeny.ai/job/sr-compiler-engineer-tensordyne-munich-wgebfvc0rdwm) — Munich, Germany
- [Staff ASIC STA Engineer](https://feeny.ai/job/staff-asic-sta-engineer-tensordyne-sunnyvale-9sppj4qyy64z) — Sunnyvale, CA
- [Simulation & Modeling Engineer](https://feeny.ai/job/simulation-modeling-engineer-tensordyne-munich-4dhthzme0wv5) — Munich, Germany
- [Director, AI Systems Solutions Engineering](https://feeny.ai/job/director-ai-systems-solutions-engineering-tensordyne-sunnyvale-2hqajfpnnvmb) — Sunnyvale, CA