Computer Vision Solutions Leader

About The Role

In this role, you will lead the design and delivery of a Computer Vision and AI Proof of Technology (PoT) engagement - from validating the technical approach to proving it with measured evidence in a live, ambiguous early-stage environment. You will work closely with the client, internal engineering teams, and business development leaders to shape a cascading, self-improving recognition architecture and translate early, evidence-based results into a scalable path to production. The role requires a mix of deep technical work and customer-facing leadership, with direct, visible impact on whether an AI-driven solution scales beyond its first proof point.

Responsibilities

• Lead end-to-end delivery of a Computer Vision PoT engagement - from discovery and solution architecture through model-cascade implementation, evidence-based evaluation, and the go/no-go decision on production scale-up
• Translate an ambiguous, evolving set of client expectations into a scoped, achievable technical roadmap, and proactively re-shape that scope as early results come in
• Act as a trusted advisor to client stakeholders, setting realistic expectations around what a PoT can and cannot prove, and communicating evidence-based results rather than overstated claims
• Lead project delivery, while directing a small cross-functional team
• Design and oversee a cascading model-training architecture (fast detection model → stronger teacher model → adjudication layer) with a learned routing rule for human review, and the evaluation harness used to measure it
• Design and implement end-to-end MLOps pipelines on Azure that automate continuous model retraining based on the model cascade approach, covering data ingestion and versioning, labeling feedback loops, training, evaluation gates, model registration, deployment, and monitoring
• Analyze the production domain, define triggers and policies for retraining cycles (e.g., drift detection, accuracy thresholds, volume of newly adjudicated labels) and ensure each cycle is reproducible, auditable, and measurable
• Guide decisions on human-in-the-loop labeling strategy, retraining cycles, and quality-control audits, balancing labeling effort against measured accuracy gains
• Support pre-sales and scale-up activities, including stakeholder workshops, readouts, proposals, and solution positioning for a subsequent production phase
• Develop reusable frameworks, proofs-of-concept, and accelerators - particularly around model cascades, edge/geospatial inference, and lightweight evaluation harnesses - to improve scalability and reduce time-to-market for similar engagements
• Establish best practices for model lifecycle management, versioned/rollback-capable model adapters, and deployment in cloud and edge environments
• Demonstrate a creative mindset and comfort operating with incomplete information, where requirements, targets, and even the underlying hardware/regulatory environment may shift during delivery
  
  

Requirements

• 6+ years of experience in Computer Vision, Machine Learning, Deep Learning, or AI solution delivery, including at least one engagement taken from early proof-of-concept through a production or scale-up decision
• Strong expertise in multimodal AI, generative AI, transformers, and vision-language models, with practical experience applying them as part of a model cascade (e.g., teacher/student distillation, adjudication, active learning)
• Hands-on experience with Python and ML frameworks such as PyTorch, Pandas, NumPy, and cloud-based AI toolsets
• Experience across the full AI model lifecycle, including experimentation, optimization, deployment, and inference, ideally including edge and post-deployment monitoring
• Strong understanding of classical Computer Vision techniques, including camera calibration, feature matching, homography estimation, geospatial/georeferenced imagery, and image/video analytics
• Experience deploying AI solutions within Azure, AWS, GCP, Kubernetes, OpenShift, or hybrid enterprise environments; familiarity with geospatial data stores (e.g., PostGIS) and/or Databricks is a plus
• Familiarity with Infrastructure as Code (e.g., Terraform, Bicep) and containerized deployment using Kubernetes (AKS) or similar orchestration platforms; experience with edge orchestration (e.g., Azure IoT Edge, K3s) is a plus
• Strong advisory and stakeholder management capabilities, with demonstrated experience setting and resetting expectations with non-technical or commercially demanding stakeholders
• Excellent communication, collaboration, analytical, and problem-solving skills, with the ability to make and defend evidence-based recommendations rather than guaranteed outcomes
• Understanding of enterprise software architecture, cloud computing, Big Data, SDLC, and the practical constraints (data availability, hardware, connectivity, regulatory) that shape AI adoption in the field
• Expertise with GPU-accelerated technologies such as DeepStream, TensorRT, Triton Inference Server, ONNX, TAO Toolkit, or similar platforms is a plus
• Bachelor’s degree in Computer Science, Applied Mathematics, Physics, or related technical field preferred
  
  

SoftServe is an equal opportunity employer. Qualified applicants will receive consideration regardless of race, color, ancestry, ethnicity, national origin, religion, sex, sexual orientation, gender identity or expression, age, citizenship, disability, health condition, marital or family status, veteran status, or any other characteristic protected by applicable law.