IBM
February 2021 — July 2025
- Designed, developed, fine-tuned and evaluated scalable NLP and machine learning models using TensorFlow, PyTorch and Hugging Face.
- Built and maintained scalable data pipelines for feature extraction, validation and preprocessing alongside data science and engineering teams.
- Architected production-grade distributed systems and microservices optimised for AI inference on Docker, Kubernetes and IBM Cloud.
- Deployed, monitored and managed models through CI/CD pipelines, ensuring reliability, automated retraining and consistent performance.
- Integrated models into existing backend architectures with seamless API interoperability.
- Addressed and mitigated model drift to keep production models stable and accurate.
- Optimised inference systems for low latency and efficient resource use across cloud and containerised infrastructure.
- Provided technical guidance, documentation and training for internal teams and external stakeholders adopting AI.