← Curriculum

C05 Stage B · Core AI 120 h (40 T / 80 P) Statutory Carries 25% of QP weight

Software Code Development for Model Deployment

Turns a model into specified, tested, documented, deployable software.

Prerequisite
F02, C01
Modules
5
NOS
SSC/N8122
Statutory mapping
Model Curriculum Module 6 — delivers SSC/N8122 in full

Tech stack

FastAPIPydanticDockerpytestlocust

Modules

5 modules · 120 hours

M1 · Define requirements

22 h · PC1, PC2
TOPICS
  • Software Requirements Specification
  • Functional and non-functional requirements: latency, throughput, availability, security, cost
  • Technical specifications and limitations of the target system
  • Evaluating the design of data flows and structures
  • Interface and schema contracts
  • Traceability matrix

Lab: Produce a complete SRS and reviewed data-flow design for a model-serving system

Course material for this module is in production.

M2 · Evaluate model and concurrency design

22 h · PC3, PC4
TOPICS
  • Reviewing core algorithmic model design for deployability
  • High-Level Design: solution and rationale, assumptions, constraints, dependencies, interfaces
  • Parallel programming constraints of the target system
  • Concurrency models
  • Thread safety
  • Statelessness and idempotency

Lab: Write and defend an HLD including a concurrency and constraint analysis

Course material for this module is in production.

M3 · Convert specifications into reusable code

34 h · PC5, PC6, PC7
TOPICS
  • Low-Level Design: programming structures, coding standards, checklists
  • Modular and reusable component design
  • Determining component reuse suitability
  • Serving with FastAPI
  • Request/response schemas with Pydantic
  • Configuration and secrets
  • Dependency injection
  • Packaging and containerisation
  • Code-generation tooling
  • Stakeholder documentation and validation walkthroughs

Lab: Implement a documented, containerised, schema-validated inference service and validate it with stakeholders

Course material for this module is in production.

M4 · Unit testing and defect management

28 h · PC8, PC9, PC10
TOPICS
  • Test design and the test pyramid
  • Unit test case creation and execution
  • Fixtures, mocking, parametrisation
  • Coverage measurement
  • Performance and load testing
  • Defect taxonomy and triage
  • Root-cause analysis
  • Corrective action
  • Recording corrective actions to improve future design
  • Refactoring for performance

Lab: Build a test suite with a coverage gate; log, classify and remediate seeded defects; optimise the hot path

Course material for this module is in production.

M5 · Approval and release

14 h · PC11
TOPICS
  • Organisational approval process for code designs
  • Pull-request and sign-off workflow
  • Release notes and versioning
  • Deployment readiness checklist
  • Handover documentation

Lab: Submit optimised, tested code through a formal approval gate to release

Course material for this module is in production.

Real-world work scenario

Data science delivers a notebook model that must go live in three weeks under a 99.5% availability target. Write the SRS and HLD, refactor the notebook into a tested, containerised, schema-validated service, load-test it, remediate the defects found, document it, and take it through code review and release approval.

Assessment

Theory 25 · Practical 75. Deliverables: SRS, HLD, LLD, running service, test suite with coverage report, defect log, approved PR.