INTERN360
// Certificate of completion
This certifies that
Baskar
@baskarmother
completed the virtual internship track
AI Agentic — Build Aria
24
tasks shipped
24
PRs merged
25
review rounds
Every task passed automated tests and code review before merging.
🎤 Understanding verified by interview assessment — 10/100
COMPLETED AUGUST 1, 2026 · REF QEDUWJA0
// Verification — the work behind this
Recorded automatically by the platform as each pull request moved through CI, AI code review, and merge. Not self-reported.
118
story points
25
commits
50
CI runs passed
25
review cycles
| Work item | Title | SP | PR | Review rounds | Merged |
|---|---|---|---|---|---|
| TASK-001 | Wire Aria to the OpenAI chat API | 2 | #1 | 1 | Aug 1 |
| STORY-002 | Assemble the chat prompt (system + history + user) | 2 | #2 | 1 | Aug 1 |
| BUG-003 | Streaming parser returns raw SSE lines instead of text | 2 | #3 | 1 | Aug 1 |
| TASK-004 | Add retries, timeout, and token-usage logging | 5 | #4 | 1 | Aug 1 |
| STORY-005 | Introduce a LangChain chat-model wrapper | 5 | #5 | 1 | Aug 1 |
| TASK-006 | Model registry — choose the model by name | 3 | #6 | 1 | Aug 1 |
| STORY-007 | Prompt templates via LangChain | 5 | #7 | 1 | Aug 1 |
| BUG-008 | Temperature/config isn't reaching the model | 2 | #8 | 1 | Aug 1 |
| STORY-009 | Build the agent with create_agent | 8 | #9 | 1 | Aug 1 |
| TASK-010 | Add a calculator tool | 3 | #10 | 2 | Aug 1 |
| STORY-011 | Add a knowledge-base lookup tool | 5 | #11 | 1 | Aug 1 |
| BUG-012 | Agent loops forever when a tool errors | 5 | #12 | 1 | Aug 1 |
| TASK-013 | Validate tool inputs | 3 | #13 | 1 | Aug 1 |
| STORY-014 | Short-term conversation memory (sliding window) | 5 | #14 | 1 | Aug 1 |
| STORY-015 | Long-term memory across sessions | 8 | #15 | 1 | Aug 1 |
| BUG-016 | Memory grows unbounded and blows the context limit | 5 | #16 | 1 | Aug 1 |
| STORY-017 | Ingest and chunk docs into embeddings | 8 | #17 | 1 | Aug 1 |
| STORY-018 | Retrieve the top-k relevant chunks | 5 | #18 | 1 | Aug 1 |
| STORY-019 | Answer with retrieved context and citations | 5 | #19 | 1 | Aug 1 |
| BUG-020 | Retrieval returns duplicates and poor ranking | 3 | #20 | 1 | Aug 1 |
| STORY-021 | Supervisor routes to specialist agents | 8 | #21 | 1 | Aug 1 |
| STORY-022 | Specialist hand-off with shared context | 5 | #22 | 1 | Aug 1 |
| STORY-023 | Expose an MCP server's tools to the agent | 8 | #23 | 1 | Aug 1 |
| STORY-024 | Evaluation harness with prompt versioning | 8 | #24 | 1 | Aug 1 |
The source repository is private to protect course material; the numbers above are recorded from GitHub webhooks (CI results, commits, merges) and the AI review pipeline. Questions? Verify via the reference code QEDUWJA0.