Software engineer, GenAI Engineering
ZakkarijaMicallef
Experience
Software Engineer, GenAI Engineering
Builds tools that speed up engineers across the company: the internal agent platform, the MCP integration platform, and the infrastructure around them.
What I built
- 01
Internal agent platform
An internal platform for AI agents, built for engineers across the company.
- 02
MCP integration platform
Connects internal tools and services to agents through the Model Context Protocol.
- 03
Developer tooling
The infrastructure and tooling around both platforms, so engineering ships faster.
Software Engineer
Java engineer on backend orchestration for a bare-metal cloud provider. Led automated RAID configuration, custom OS image creation, and internal provisioning tools built with Spring Boot.
What I built
- 01
Backend orchestration
Java backend orchestration for a bare-metal cloud provider.
- 02
Automated RAID configuration
Led the project to configure server RAID automatically.
- 03
Custom OS images
Led the work on creating custom operating-system images.
- 04
Provisioning tools
Internal tools for provisioning servers, built with Spring Boot.
Software Engineer Intern
Part of a small intern team that built and ran an internal employee-management tool end to end: frontend, backend, database, deployment and support. Worked with Product Owners on sprint priorities. Java, Spring, Maven, SQL and JavaScript.
What I built
- 01
Employee-management tool
An internal tool, built and run end to end by a small intern team.
- 02
Full stack, start to finish
Frontend, backend, database, deployment and support, with sprint priorities set alongside Product Owners.
Computer science and artificial intelligence
- 2023–2025
M.Sc Computer Science
Thesis with : industrial anomaly-detection pipelines on CNC machine signals, built once on MLflow and once on Kubeflow to compare the two in practice.
- 2018–2021
B.Sc Artificial Intelligence
Machine learning, computer vision and knowledge representation. Dissertation on saliency-directed product placement.
Projects
Peer-reviewed paper, 2026
A Systematic Review of MLOps Tools
Maps MLOps-native tools to lifecycle components, across 41 papers, and synthesises the benefits and limitations reported from real use. Orchestration, data versioning, experiment tracking and managed cloud platforms dominate. No single tool covers the whole lifecycle, so teams stitch several together and interoperability becomes the central problem.
MLflow gets a team running quickly; Kubeflow needs Kubernetes fluency before it gives anything back.
MSc thesis, 2025
Industrial MLOps for anomaly detection
Two pipelines for the same industrial anomaly-detection use case on CNC machine signals, with : one built on MLflow, one on Kubeflow. MLflow gets a team running quickly; Kubeflow needs Kubernetes fluency before it gives anything back. Which to pick depends mostly on what the team already runs.
BSc dissertation, 2021
Saliency-directed product placement
Ranks the products in a scene by how likely each is to catch attention first, combining Mask R-CNN object detection with a saliency-segment ranking algorithm. The ranking reached a 0.66 correlation with human attention patterns.
Contact
Get in touch
zak.micallef27@gmail.comOpen to new opportunities, including small contract projects.