Key Responsibilities :

1. Incident Management & Bug Fixing :

- Investigate and resolve incidents reported on AI algorithms (model crashes, wrong outputs, data pipeline failures, performance degradation).

- Diagnose root causes: data issues, code bugs, dependency conflicts, model drift.

- Apply fixes, test and validate corrections before redeployment.

- Document incidents and resolutions in the ticketing system.

2. Algorithm Maintenance :

- Apply patches, dependency updates and environment fixes.

- Ensure models remain functional after infrastructure changes (cloud updates, library upgrades, data schema changes).

- Maintain and update data preprocessing pipelines feeding the models.

3. Code Review & Quality :

- Read, understand and improve existing Python codebases (often written by Data Scientists).

- Refactor poorly structured code to improve maintainability and reliability.

Experience :

- 3+ years of experience.

Core Technical Skills :

- Python: Advanced Python, debugging, OOP, clean code (Expert).

- Data manipulation: Pandas, NumPy, data wrangling, ETL pipelines (Intermediate).

- ML Frameworks: Scikit-learn, TensorFlow or PyTorch (reading & fixing, not training from scratch) (Intermediate).

- API & services: FastAPI, REST APIs, JSON, basic microservices (Intermediate).

- Version control: Git + GitHub (Advanced) (Intermediate).

- Environment management: Poetry, pip, virtual environments, requirements management (Expert).

- Cloud - GCP AI Services: GCP Projects, Vertex AI, BigQuery, Service Account, Google Storage (Buckets), Artifact Registry (Beginner).

- Containerisation: Docker - run, build, debug a container (Intermediate).

- SQL: Ability to query databases and investigate data issues (Intermediate).

GitHub - Advanced Skills:

- The consultant must demonstrate advanced proficiency with GitHub beyond basic commit/push usage, including:

1. Repository & Code Management: Branching strategies (GitFlow, trunk-based development, feature branches), Pull Request lifecycle (creating, reviewing, requesting changes, approving and merging), Resolving merge conflicts confidently on complex Python codebases, Tagging, releases and semantic versioning.

2. Collaboration & Governance: Branch protection rules (required reviews, status checks before merge), Code review best practices (inline comments, suggesting changes directly in PRs).

3. Advanced Features: GitHub Packages (publishing and consuming internal Python packages), Pre-commit hooks and linting integration (Black, Flake8, isort), Git history investigation (git bisect, git blame, git log - graph to trace the origin of a bug).