Junsheng

Mapped the recommendation system, tracing input and output flows to find gaps in the pipeline.

Explored MLOps practices and evaluated the production system against best practices.

Independently scoped, built and presented a prototype to address a key gap.

Built workflow automations for European SME clients, from prototyping through deployment.

Researched AI/ML use cases for energy-sector SaaS and IoT gateway products.

Developed dashboards, analytics and internal tools while contributing to the SaaS platform using AI-assisted development tools.

Building freight-workflow agents for traders and charterers, with containerised APIs and health checks across environments.

Developed MCPs for chartering analysers, reducing query time from 120s to 10s through better prompts and context efficiency.

Fine-tuning internal models for more efficient responses and more accurate internal context.