Software Developer

About the Role

We are looking for a Software Developer with a strong foundation in deep learning to join the Software Lab in Lucknow. You will turn state-of-the-art multimodal models into production-grade solutions on the IBM watsonx platform, covering defect and anomaly detection, segmentation, visual inspection and document understanding for clients in energy, infrastructure, logistics, BFSI and the public sector. You will design and build Python services and data pipelines and fine-tune foundation models on client data, and package them as containerised, API-driven services on Red Hat OpenShift and IBM Cloud. You will also make models robust and explainable for real-world conditions, and work with pre-sales teams to build proofs of concept and present results to clients. You will be a self-starter, requiring limited supervision or guidance, with good communication skills and the ability to multi-task and prioritize your work. Working with researchers, architects and client stakeholders will be critical to taking ideas from prototype to production.

Key Responsibilities

• Design and build Python services and data pipelines.
• Fine-tune foundation models on client data.
• Package models as containerised, API-driven services on Red Hat OpenShift and IBM Cloud.
• Make models robust and explainable for real-world conditions.
• Work with pre-sales teams to build proofs of concept and present results to clients.

Requirements & Tech Stack

• Python
• Deep Learning
• PyTorch or TensorFlow
• CNN, Transformer, Detection and Segmentation architectures
• Computer Vision tasks (anomaly detection, defect detection, image segmentation, classification)
• Building data pipelines for large image datasets (cleaning, augmentation, annotation workflows)
• Model evaluation, statistical analysis and error analysis
• Building API-driven services (REST / FastAPI)
• Packaging models with Docker
• GitHub as source code control

Qualifications

• Doctorate Degree (Required Education)
• Master's degree in Computer Science, AI, Robotics or equivalent (Required Technical and Professional Expertise)

Preferred Qualifications

• PhD (completed or in progress) in AI, Computer Vision or Robotics
• Publications at peer-reviewed venues such as CVPR, WACV, BMVC or ACM SAC
• Experience with vision-language models (e.g. CLIP, SAM) and parameter-efficient fine-tuning (adapters, LoRA)
• Experience with industrial inspection, infrastructure monitoring or medical imaging use cases
• Knowledge of explainable AI techniques such as Grad-CAM, and adversarial robustness
• Knowledge of model optimisation and edge deployment (ONNX, TensorRT, OpenVINO)
• Knowledge of IBM watsonx, Hugging Face, Kubernetes, Docker and Red Hat OpenShift
• Knowledge of any cloud services, including IBM Cloud, AWS, Azure and GCP
• Experience with Linux systems
• Experience with C++, Java or other languages (useful)