Responsibilities
- Shape and lead partner-driven research projects, translating food security, public health, climate resilience and environmental priorities into rigorous research questions, measurable technical objectives and decision-support requirements.
- Develop and adapt computer vision, geospatial machine learning and foundation models for classification, segmentation, change detection and spatiotemporal forecasting, selecting approaches appropriate to each problem and available data.
- Build reproducible data pipelines that integrate satellite imagery, raster and vector layers, field observations and sector statistics, with documented provenance, quality checks and appropriate data-access controls.
- Evaluate models using suitable baselines, spatial and temporal validation, uncertainty estimates and independent reference data; assess sampling bias, transferability and performance under limited-data conditions.
- Lead the technical transition from research prototypes to usable workflows and tools with engineers and partners, considering computational cost, connectivity, usability, monitoring and long-term maintenance.
- Co-develop and validate solutions with government, academic and nonprofit collaborators, incorporating domain expertise, field feedback and responsible AI, privacy and security practices.
- Lead and contribute to peer-reviewed publications; share methods, datasets or code where permitted, and communicate results, assumptions and limitations clearly to technical, operational and policy audiences.
- Mentor interns and researchers, guide scientific reviews and technical decisions, and contribute to an inclusive research community grounded in Microsoft's values of respect, integrity and accountability.
Qualifications
Required Qualifications:
- Doctorate in computer science, machine learning, statistics, remote sensing, geospatial science, electrical engineering or a related field, OR equivalent experience.
- At least two years of experience developing and evaluating deep learning models, including applications using geospatial or Earth-observation data.
- At least two years of experience with modelling in Python, including use of PyTorch or TensorFlow, with experience processing raster and vector data and building reproducible research code.
- Experience designing experiments, applying statistical methods and assessing model performance, with evidence of communicating research findings and collaborating across technical and domain teams.
Preferred qualification:
- Two or more peer-reviewed publications or conference presentations in machine learning, computer vision, remote sensing, geospatial AI or related applied research.
- Two or more years of experience delivering applied AI projects in areas such as food and water security, public health, disaster response, climate resilience or environmental conservation.
- Experience with geospatial foundation models, multispectral or radar imagery, multimodal or self-supervised learning, or applications involving bioacoustics or language models.
- Experience providing technical leadership and mentoring, and co-developing cloud-based tools with public-sector, academic or nonprofit partners for adoption in resource-constrained environments.