Josephine Sullivan

The Division of Robotics, Perception and Learning at KTH Royal Institute of Technology invites applications for a doctoral student position focused on image representations for class discovery. This WASP-funded project aims to advance generative approaches for out-of-distribution discovery and novel species identification, with a particular emphasis on fine-grained classification tasks such as plankton species identification. The research will integrate genomic measurements into the identification process, developing robust algorithms that are application area independent but with environmental and biological relevance. The successful candidate will join a collaborative environment involving Associate Professor Josephine Sullivan (RPL), Professor Anders Andersson (Environmental Genomics, SciLife Lab), and Bengt Karlson (SMHI). The PhD student will also be part of the WASP graduate school, benefiting from interdisciplinary expertise and scientific output, as highlighted in the article "The ocean's smallest creature is mapped." The project offers opportunities to contribute to cutting-edge research in computer vision, deep generative learning, and environmental genomics. Applicants must meet the eligibility requirements for postgraduate education as outlined by the Swedish Higher Education Ordinance. This includes holding a second cycle degree (such as a master's) or equivalent, or completing at least 240 higher education credits with 60 at the second-cycle level. Practical proficiency in deep learning programming libraries (TensorFlow, PyTorch, JAX) is mandatory, and experience with GPU-based experimentation and cluster computing (Docker, Slurm) is advantageous. English proficiency equivalent to English B/6 is required. Selection criteria include academic achievements, completed courses, demonstrated programming ability, and personal skills such as independence, collaboration, professionalism, and analytical thinking. Specialization in computer vision and/or machine learning is highly desirable. The position is full-time, temporary, and offers a monthly salary according to KTH's doctoral student salary agreement. Employment is for up to four years, with renewal options, and includes employee benefits and a supportive workplace. The doctoral student may perform limited additional tasks (up to 20%) related to training and administration. The position is based in Stockholm, Sweden, and may be subject to security clearance if classified as security-sensitive. To apply, candidates must submit a complete application through KTH's recruitment system, including certified copies of diplomas, grades, proof of language requirements, CV, and relevant publications or technical reports. Applications must be received by midnight CET on the closing date. For further information, contact Associate Professor Josephine Sullivan at [email protected]. KTH Royal Institute of Technology is a leading international technical university committed to education, research, and innovation for a sustainable society. The university values equality, diversity, and equal opportunities, offering a creative and dynamic environment for personal and professional growth.

Doctoral Student in Urban Heat Prediction Using AI and Earth Observation at KTH Royal Institute of Technology in Stockholm, Sweden, is an open PhD position in the third-cycle subject Computer Science . The project sits in the Division of Robotics, Perception and Learning and is focused on a timely and socially important problem: mapping urban air temperature at high spatial and temporal resolution. Cities are increasingly affected by urban heat, including short spikes and longer heat periods, but dense high-resolution temperature mapping remains challenging. The research will explore the use of conditional diffusion models and other deep generative learning approaches to fuse heterogeneous data sources such as satellite observations, forecast data, and crowd-sourced station measurements into probabilistic urban temperature maps. This is a cutting-edge machine learning problem with clear relevance to climate resilience, urban environments, and earth observation. The doctoral student will join an active research collaboration involving Associate Professor Josephine Sullivan at RPL, Professor Yifang Ban from the Division of Geoinformatics at the Department of Urban Planning and Environment, and Sebastian Hafner at RISE in Kista. Supervision is proposed by Associate Professor Josephine Sullivan . Applicants should have strong preparation for postgraduate study, including a relevant second-cycle degree or equivalent academic background. The position requires practical proficiency in deep learning and demonstrated programming ability, particularly with libraries such as TensorFlow , PyTorch , or JAX . Experience with GPU-based experimentation and cluster computing tools such as Docker and Slurm is advantageous. English proficiency equivalent to English B/6 is mandatory. Prior experience in machine learning, computer vision, and/or remote sensing is highly desirable. Funding is provided through Digital Futures , and the doctoral student receives a monthly salary according to KTH’s doctoral student salary agreement. The appointment is a full-time temporary doctoral position. Applications must be submitted through KTH’s recruitment system by 2026-09-24 . Application materials include diplomas and transcripts, evidence of language qualifications, a CV, a motivation letter describing research interests and goals, and representative publications or technical reports. KTH asks applicants to ensure that submissions are complete and follows the stated instructions carefully. This position is a strong fit for candidates interested in AI for environmental sensing , remote sensing , computer vision , urban climate modeling , and probabilistic machine learning within a leading technical university environment in Sweden.