Postdoctoral Fellow in Machine Learning and Computer Vision (Digital Agriculture)

This role focuses on solving complex semantic segmentation and row detection challenges by developing custom unsupervised domain adaptation (UDA) pipelines and generative deep learning architectures.

The successful candidate will work at the intersection of foundational machine learning theory and large-scale industrial deployment, bridging the gap between cutting-edge research and the practical needs of the agriculture sector through our partnership with MacDon Industries.

Key Responsibilities
• Research & Model Development: Design, train, and optimize state-of-the-art neural network architectures, including UDA pipelines, generative models (GANs, Diffusion), and Transformers tailored for agricultural environments.
• System Implementation: Build and maintain robust ML pipelines, incorporating automated regression testing to ensure reproducibility.
• Data Leadership: Preprocess and analyze complex agricultural datasets, ensuring high-quality readiness for deep learning applications.
• Academic Output: Conduct comprehensive literature reviews, draft technical reports, and co-author publications in Q1/Q2 journals or CORE A/A* conferences.
• Mentorship: Provide technical guidance and supervision to undergraduate and graduate students (BSc/MSc level).

Required Qualifications
• Education: PhD in Computer Science or a closely related field with formal training in Machine Learning, Deep Learning, or Computational Agriculture.
• Research Record: Proven track record as a first or second author in Q1/Q2 peer-reviewed journals (Image Processing/CV) or publications in CORE A/A* conferences.
• Professional Experience: * Minimum of 2 years of applied ML experience (research or industry) with minimal supervision.
o Hands-on internship experience in Computer Vision (CV), ideally within the digital agriculture sector.
• Frameworks & Tools: Proficiency in Python and deep learning frameworks (PyTorch preferred, TensorFlow, or JAX). Experience with Linux-based environments and Docker is required.

Technical Proficiencies
Candidates should demonstrate solid knowledge or familiarity with:
• Architectures: CNNs (UNet, VGG), Semantic Segmentation (SegFormer), Transformers, and ViTs.
• Generative & Adaptation: GANs, VAEs, Diffusion Models, and Unsupervised Domain Adaptation (UDA).
• Object Detection: YOLO architectures.
• Optimization: At least two of: Quantization, Distillation, Pruning, or LoRA.
• Emerging Tech: Awareness of RAG/CoT for LLMs, State-Space Models, GNNs, and Vision-Language Models (VLMs).

Skills & Abilities
• Communication: Excellent oral and written skills (minimum CLB 10+ equivalent for non-Canadians).
• Organization: Strong time management skills and the ability to meet strict grant-driven deadlines.
• Collaboration: A proven ability to work effectively within a multidisciplinary team.

Apply through UM Careers https://viprecprod.ad.umanitoba.ca/default.aspx?REQ_ID=45466&Language=1

To help us track our recruitment effort, please indicate in your cover/motivation letter where (jobs-near-me.eu) you saw this job posting.