---
title: 'Senior AI Researcher- Pre-training (f/m/d) at Aleph Alpha'
canonical: 'https://feeny.ai/job/senior-ai-researcher-pre-training-f-m-d-aleph-alpha-heidelberg-s5v2ewqeev20'
type: 'job'
last_seen: '2026-10-18'
---

# Senior AI Researcher- Pre-training (f/m/d) at Aleph Alpha

- **Company:** Aleph Alpha
- **Location:** Heidelberg, Germany
- **Employment:** full-time
- **Work type:** hybrid
- **Posted:** 2026-10-09
- **Last confirmed live:** 2026-10-18
- **Apply:** https://jobs.ashbyhq.com/AlephAlpha/80dc6ee6-dece-4d4b-9906-ea930338f245/application

**Skills:** Python, PyTorch, Machine Learning, Software Engineering, Transformer Training Dynamics, Distributed Training, Model Architecture Design, Experiment Design, Large Language Models, Multimodal Models, torchtitan, Megatron-LM, DeepSpeed, Sparse Training, Mixture-of-Experts

> The Senior AI Researcher will advance the architecture and training of foundation models for large-scale pre-training. Responsibilities include optimizing training recipes, developing scaling strategies, designing PyTorch architectures, and ensuring training stability and efficiency on large GPU clusters.

## Job description

## Our Mission

Aleph Alpha is one of the few companies in Europe doing serious foundation model pre-training. Our customers — in finance, manufacturing, and public administration — need models that understand German, meet European regulatory requirements, and work reliably in high-stakes settings. We’re building that in Heidelberg. We are hiring a Senior AI Researcher to join our Pre-training team and to advance the architecture and training of our next generation of foundation models. If you are excited about designing inference-efficient architectures, optimising training recipes that scale reliably, and training models on a large scale cluster (thousands of NVIDIA Blackwell GPUs), we would love to hear from you.

## Team Culture

We foster a culture built on ownership, autonomy, and empowerment. Teams and individual contributors are trusted to take responsibility for their work and drive meaningful impact. We maintain a flat organisational structure with efficient, supportive management that enables quick decision-making, open communication, and a strong sense of shared purpose. We collaborate closely on complex technical problems, working in pairs or using mob programming to resolve challenging issues.

## About the Role

As a Senior AI Researcher in Pre-training (f/m/d), you will own the critical technical levers that determine the success of our next-generation models: architecture, optimization, stability, and scaling. Working at the high-leverage intersection of research and engineering, you will translate mathematical reasoning and empirical observations into principled training decisions - from small-scale proxy experiments to multi-thousand-GPU runs. We are looking for an expert who can combine rigorous experimental design with high-quality production code, directly influencing model quality, run reliability, and the efficiency of the models we ship.

## Your Responsibilities

- Recipe & Architecture Optimization: Own core elements of the training recipe (optimizers, schedules, initialization) and design PyTorch-based architectural improvements to maximize convergence, stability, and training efficiency.
- Scaling Strategy & Predictability: Develop hyperparameter scaling laws and scale-up methodologies, using small-scale proxy experiments to reliably predict multi-thousand-GPU behavior and de-risk major training decisions.
- Stability, Diagnostics & Debugging: Investigate complex convergence issues (loss spikes, divergence) and resolve hard-to-reproduce distributed system failures like communication bottlenecks, race conditions, and synchronization errors.
- System-Model Co-Design: Partner with Compute Performance, Data, Evaluation, and Post-Training teams to align the model lifecycle with hardware constraints, memory bandwidth, and communication topologies.

Core Qualifications

- You are proficient in Python and deeply familiar with PyTorch-based training workflows.
- You have a strong track record in machine learning research and software engineering, demonstrated through shipped models, impactful open-source contributions, or published research.
- You have a strong mathematical foundation and are comfortable reasoning formally about optimisation, scaling behaviour, and training dynamics.
- You deeply understand transformer training dynamics, optimisation, and the behaviour of large distributed training jobs.
- You can design rigorous experiments, reason clearly from noisy results, and translate empirical observations into robust training decisions.
- Hands-on experience pre-training large models (e.g., 7B+ parameters) on substantial infrastructure (e.g., 100+ GPU clusters).
- You apply strong software engineering practices, including writing maintainable, well-tested code and supporting reproducible experimentation workflows.
- You are able to implement complex model architectures efficiently and reliably and to debug complex issues across model code, training dynamics, and distributed systems.
- You collaborate effectively within a research and engineering team and communicate clearly about your work across Pre-training and the broader AAR/AA organization.
- You are able to work in Germany and collaborate regularly on site in Heidelberg as part of the Pre-training team.

## Preferred Qualifications

- Large-Scale Training: Hands-on experience training LLMs or multimodal models on large GPU clusters using distributed frameworks (e.g., Megatron-LM, DeepSpeed, torchtitan).
- Predictive Scaling: Familiarity with scaling laws, hyperparameter transfer, or methods for predicting large-scale training behavior from smaller proxy runs.
- Stability & Performance: Experience profiling distributed jobs and diagnosing training anomalies like loss spikes, numerical instability, or optimizer pathologies.
- Advanced Architectures: Exposure to sparse training approaches (e.g., Mixture-of-Experts) and an understanding of their routing and systems trade-offs.
- Track Record of Impact: Demonstrated research excellence through top-tier publications (NeurIPS, ICML, ICLR), impactful open-source contributions, or significant shipped technical work.
- Systems Curiosity: Low-level kernel optimization is not required, but we highly value a strong curiosity about the hardware and systems constraints that shape scale.

## What we offer

- Become part of an AI revolution!
- 30 days of paid vacation
- Access to a variety of fitness & wellness offerings via [Wellhub](https://wellhub.com/de-de/)
- Mental health support through [nilo.health](http://nilo.health)
- Substantially subsidized company pension plan for your future security
- Subsidized Germany-wide transportation ticket
- Budget for additional technical equipment
- Flexible working hours for better work-life balance and hybrid working model
- [JobRad®](https://www.jobrad.org/) Bike Lease

## About Aleph Alpha

## Company Overview
- **One-liner**: Aleph Alpha is a German AI startup that develops sovereign large language models (SLLMs) for enterprises and governments, emphasizing data sovereignty, regulatory compliance, and domain-specific performance.
- **Entity Type**: Private (Series B; acquired by Cohere in a landmark agreement)
- **Headquarters**: Heidelberg, Baden-Württemberg, Germany
- **Founded**: 2019
- **Founders**: Jonas Andrulis, Samuel Weinbach

## Core Business
- **Primary industry/industries**: Artificial Intelligence (Generative AI, Large Language Models), Enterprise Software, Government Technology
- **Target customers**: B2B exclusively — enterprises in industry, public sector, and defense where data sovereignty and domain expertise are critical
- **Mission or purpose statement**: "We empower enterprises and public institutions to transform their most critical knowledge." The company operates under the principles of Trust, Responsibility, and Sovereignty.

## Products & Services
- **Luminous LLM**: A family of large language models designed for enterprise and government use, with a focus on transparency of sources used to generate results. The models are trained on customer-specific domains (legal, administrative, industrial, scientific) and run exclusively on European infrastructure.
- **Sovereign Large Language Models (SLLMs)**: Custom-built, secure language models tailored to a customer's domain and workflows, paired with custom evaluation frameworks to ensure performance before deployment.
- **AI Agent Solutions**: Deployed for tasks like retrieving verifiable insights from sensitive documentation (e.g., reducing search time by 90% at a global chip manufacturer) and AI-powered requirements engineering (e.g., 40% time savings on RFQ processing for an automotive supplier).
- **AI Assistant for Government**: A sovereign AI assistant deployed in a government agency, empowering 80,000 users to streamline administrative processes.

## Market Standing
- **Valuation/Market Cap**: Not publicly disclosed (private company)
- **Key Metric**: Annual Revenue of $100M (as of latest data); Total Funding of $633.6M (across 6 rounds, including a $500M Series B led by Schwarz Group, Bosch Ventures, and others)
- **Notable Investors/Partners**: Cohere (acquirer), Schwarz Group, Bosch Ventures, 468 Capital, LEA Partners, Lakestar, EarlyBird, and 18+ other investors. Key partners include global chip manufacturers, automotive suppliers, and government agencies.
- **Growth Signals**: Headcount of 275 employees (+17% YoY); operates in 9 countries (Germany, Switzerland, Romania, Nigeria, Brazil, France, Peru, Austria, Netherlands); acquired thingsTHINKING (April 2025); landmark agreement with Cohere for operational realignment and partnership.

## Competitive Advantages
- **Data Sovereignty & Compliance**: All models run on European infrastructure and are trained in accordance with EU law, making them uniquely suited for regulated industries (defense, public sector, legal) where data cannot leave the continent.
- **Domain-Specific Customization**: Unlike general-purpose models, Aleph Alpha's SLLMs are trained on a customer's own domain knowledge and workflows, with custom evaluation frameworks that prove performance before deployment.
- **Proven Track Record in Sensitive Environments**: Demonstrated success in government (80,000 users), defense, and industrial settings where precision and reliability are non-negotiable.
- **End-to-End Partnership Model**: The company co-builds solutions with customers from first use case through production deployment, ensuring the technology fits existing processes.

## Strategic Focus
- **Current priorities**: Deepening the partnership with Cohere for operational realignment and scaling; expanding sovereign AI solutions for European enterprises and governments; continuing to build domain-specific SLLMs for legal, administrative, industrial, and scientific sectors.
- **Direction for growth**: Focus on mission-critical, sensitive environments where general-purpose AI fails; expanding across European public sector and defense; leveraging the Cohere partnership for broader reach and resources.

## Why Work Here
- **Culture highlights**: The company emphasizes collaboration, openness, and a deep commitment to responsibility. The team is described as international, with engineers, scientists, and innovators working across four German locations. The mission is centered on "Trust. Responsibility. Sovereignty."
- **Remote/hybrid/office policy**: Based on job postings, roles are primarily on-site in Heidelberg, with some flexibility. The company has locations in Heidelberg, Speyer, and other German cities.
- **Notable perks or engineering culture**: Strong focus on research and engineering (33% of workforce in technical roles, 8% in research). The tech stack includes TensorFlow, Kubernetes, Flask, MLflow, and more. The company has filed 4 patents related to AI model compression and performance. Recent departures of key executives (e.g., Co-CEO Ilhan Scheer, COO Ben Oppl) suggest a period of transition, but the company is actively hiring for senior AI research and engineering roles.
- **Employee ratings**: 3.2/5.0 on employer review platforms (Work-Life: 3.5, Compensation: 3.5, Culture: 3.1, Career: 3.5). The company has 15 active job postings, with a strong focus on senior AI researchers and software engineers.

## Sources
1. [Aleph Alpha Website](https://aleph-alpha.com/)
2. [Aleph Alpha Careers Page](https://jobs.ashbyhq.com/AlephAlpha)
3. [LinkedIn Company Profile](https://www.linkedin.com/company/aleph-alpha/)
4. [CB Insights Company Profile](https://www.cbinsights.com/company/aleph-alpha)

## Other roles at Aleph Alpha

- [Senior AI R&D Engineer - Performance - Pre-training (f/m/d)](https://feeny.ai/job/senior-ai-r-d-engineer-performance-pre-training-f-m-d-aleph-alpha-heidelberg-63c77hha52h8) — Heidelberg, Germany
- [Senior AI Researcher: LLM post-training for hallucination reduction (f/m/d)](https://feeny.ai/job/senior-ai-researcher-llm-post-training-for-hallucination-reduction-f-m-d-aleph-zap2g8gcb5g4) — Heidelberg, Germany
- [Senior AI R&D Engineer – Model Training  (f/m/d)](https://feeny.ai/job/senior-ai-r-d-engineer-model-training-f-m-d-aleph-alpha-heidelberg-xqr3y48gda71) — Heidelberg, Germany