Omegga · Munich ·
Build and improve deep-learning models for spectroscopy-based AI that detects chick sex in eggs, iterating from research to robust deployment in noisy real-world settings.
atoms · San Francisco, CA
Tools for Humanity · Munich
absentia-labs · Boston
2K · Dublin, Ireland
At Omegga, we're on a mission to reinvent how AI and spectroscopy can drive positive change for animals, people and the planet. We started with a clear objective: eliminate chick culling on a global scale through early, non-invasive in-ovo sex detection. This outdated practice still impacts billions of lives, and we're here to change that. We're not building for the niche, we're building for the global standard.
Today, our systems are already in use, and the focus is on scaling performance, robustness, and deployment across an international customer base. For us, scaling means more than growth. It means changing an industry for the better.
We are looking for people who take ownership, think in systems, and want to help turn a working product into the industry benchmark. If that sounds like you, we'd love to hear from you.
We are looking for a Deep Learning Engineer with strong research depth and hands-on implementation skills to push our model performance forward.
You will work on deep learning architectures and iterate rapidly: from hypothesis and experiments to evaluation, monitoring, and integration into our algorithm stack. Your work will directly improve detection quality and robustness in a real-world, noisy environment.
Advance our deep learning models through research iteration: develop and test architectural improvements, training objectives, and optimization techniques. You will own the loop from idea → experiment → conclusion → next iteration.
Build rigorous evaluation and benchmarks: define evaluation sets, establish clear metrics (precision, recall, accuracy, calibration), and create repeatable benchmark runs so improvements are measurable and comparable over time.
Own monitoring of model quality: set up monitoring for model performance and data shifts, define alerting signals, and build lightweight reporting that makes regressions visible early.
Partner cross-functionally to turn findings into impact: work with data/engineering teams to improve datasets and labeling strategies, and with product/ops stakeholders to align on what “good” looks like in practice.
MSc in Computer Science or a related field with 4+ years of applied deep learning experience, or a PhD — paired with a proven track record of taking research from idea to working system.
Strong understanding of optimizing modern neural architectures, with demonstrated depth in transformer architectures and practical experience adapting them to specific domains (e.g., attention variants, efficiency improvements, robustness, training stability).
Ability to architect problem-specific models — adapt modern architectures, losses, and training objectives to the structure of the task, treating published work as a starting point rather than reaching for an off-the-shelf model
Strong Python deep learning stack experience (e.g. PyTorch, Tensorflow), including training pipelines, experimentation discipline, and reproducibility.
Real depth in the mathematical core of modern ML: probability, optimization, information theory, statistical inference. You can derive a loss function rather than just import one.
Solid experience with experiment tracking and model evaluation tooling (e.g. Weights & Biases or similar), and a strong bias for measurement-driven progress.
Fluency in English, German is a plus
Experience with domain adaptation and evaluation in imbalanced settings (rare events, high cost of misses/false alarms).
Familiarity with deployment-adjacent concerns: model packaging, performance constraints, and continuous evaluation in changing real-world conditions.
Experience working with sensor/time-series or industrial data, where edge cases and dataset shifts are the norm.
Experience with agentic AI development workflows to speed up experimentation (analysis, ablations, test scaffolding) while maintaining careful review and scientific rigor.
Proven ability to deliver in fast-paced, high-ambiguity environments.
Mission & Technical Challenge: At Omegga, you work on problems that are both technically demanding and globally relevant. We build deep-tech solutions with real-world impact, pushing the boundaries of what's currently possible.
Relocation Support: Moving to Munich? We help you find an apartment in the Werksviertel within walking distance of the office. We connect you with the right contacts and take the stress out of apartment hunting.
Work Environment: Located in Munich's Werksviertel, we combine a focused, high-performance culture with flexibility. Up to 50% remote work, depending on role and execution needs.
Ownership: VSOPs (Virtual Stock Options) so you share in the value you help create.
Time Off: 28 vacation days per year, plus December 24th and 31st off.
Monthly meal budget for whatever keeps you going, from quick lunches to proper team dinners
Monthly perks budget to spend where it actually matters to you — fitness, mobility, or the occasional "I earned this" purchase
Annual learning budget to invest in your growth, from conferences to deep dives into new domains
Free drinks, coffee, and snacks in the office to keep you focused throughout the day
Find more English Speaking Jobs in Germany on Arbeitnow
Reddit · Remote - United States