L
Machine Learning & AIAI relevant
Company
liveeo-gmbh
Location
Berlin, Germany
Employment type
Full-time
Seniority
Mid level
Primary category
Machine Learning & AI
Posted date
23 Feb 2026
Valid through
24 Apr 2026
We are looking for a Senior ML Engineer (f/m/x) to build and scale multitemporal, multimodal computer vision models for Earth observation, combining very high resolution optical and Synthetic Aperture Radar (SAR) data into robust representations that enable semantic understanding across sensors and time. This is a balanced role: part applied research, part engineering, all impact. You’ll work across the full ML R&D lifecycle: data standardization and preprocessing, model development and training at scale, and rigorous evaluation—with a clear path from experimentation to production-grade capability.
LiveEO is a young, dynamic team that thrives on big challenges and fast learning cycles—we move quickly, stay curious, and genuinely enjoy building together. We’re on a mission to break the “curse of Earth Observation”: turning incredible satellite data into reliable, actionable decisions that people can trust and use in real operations. In this role, you’ll work in a fun, high-ownership environment where ambitious technical problems (multimodal SAR/optical foundation models) meet real-world impact—and where your ideas can go from whiteboard to production in tight, collaborative iterations.
Tech stack & tools, which potential candidate will work with:
Python (core ML development)
PyTorch + PyTorch Lightning (model training, experimentation)
Databricks + MLflow (experiment tracking, model registry)
Ray (distributed compute)
Prefect (workflow orchestration)
AWS (cloud infrastructure)
Geospatial stack: GDAL, Rasterio, GeoPandas, STAC (EO data handling and standardization)
Datastores: PostgreSQL (metadata / operational data)
As a Senior ML Engineer -Remote Sensing & Foundation Models (f/x/x), you will drive the development of state-of-the-art ML systems that can learn from and reason about large volumes of satellite imagery.
Identify and adapt SOTA approaches in remote sensing and foundation models (papers → prototypes → validated baselines), focusing on pragmatic wins under real constraints.
Design, train, and iterate on bitemporal and multimodal SAR–optical models (alignment/fusion, robust embeddings, bitemporal/multitemporal representations), with clear ablations and measurable performance improvements.
Own EO data standardization & preprocessing for high resolution SAR and optical imagery (normalization/calibration choices, tiling/chipping, pairing/co-registration sanity checks, sampling/augmentations) and drive dataset quality diagnostics.
Build scalable training + evaluation pipelines in our stack (Databricks, PyTorch Lightning, MLflow), including experiment tracking, reproducibility, and systematic failure analysis across geographies and acquisition conditions.
Deliver production-ready ML components (robust inference interfaces, model packaging, deterministic evaluation, monitoring signals/model cards) that downstream teams can depend on.
Collaborate closely with product teams to ensure the models translate into business value and with the data annotation team to define labeling guidelines and close feedback loops on edge cases and quality.
Must have:
Strong Python engineering fundamentals with clean, maintainable coding style.
Deep experience with PyTorch and PyTorch Lightning.
Experience implementing and training deep learning models at scale.
Strong understanding of ML experimentation, versioning, and tracking via MLflow and Databricks.
Strong CV fundamentals (representation learning, supervision strategies, evaluation design) and practical debugging/optimization skills.
Hands-on experience with satellite imagery; strong preference for experience spanning optical and SAR.
You take ownership and proactively push work forward.
You communicate clearly and collaborate smoothly both within and across teams.
Pragmatic mindset: You balance deep research with practical delivery.
You enjoy working with complexity and turning ambiguity into structure.
Living in or are willing to relocate to Berlin
Nice to have:
Experience with Ray for distributed computing.
Experience with Prefect (or similar orchestration) for ML workflows.
Familiarity with AWS or other cloud platforms.
Geospatial experience including GDAL, Rasterio, GeoPandas, STAC, and basic SAR preprocessing libraries.
Knowledge of PostgreSQL (or similar).
Familiarity with Vision-Language Models (VLMs) and/or LLMs (e.g., CLIP-style contrastive learning, multimodal finetuning, prompt/instruction tuning for vision-language).
Experience pretraining or adapting large-scale geospatial foundation models (self-supervised learning, contrastive objectives, masked modeling, retrieval-based evaluation).