Week 08 — Evals II in CI + observability — interview-2¶

At a glance

Phase 2: RAG + Evals · Starts: Mon 16 Nov 2026 · Planned: 16h 45m

Build: Capstone: promptfoo/DeepEval/Ragas in CI with thresholds + Langfuse tracing

Interview checkpoint: interview-2

Run the full mock loop and score it with the rubric.

By day¶

System Design · 45 min · Search systems: inverted index, BM25, sharding an index, near-real-time refresh → Search systems & inverted indexes w08-sd-1
DSA · 25 min · Invert Binary Tree + Maximum Depth of Binary Tree → Trees: DFS, BFS, BST · resource w08-dsa-1
Python · 20 min · Rep: Polars/DuckDB to aggregate eval results (per failure category, per model) → Data tooling: Polars, DuckDB, Arrow w08-py-1
Communication · 30 min · Grammar: Common errors of fluent speakers: uncountables, word order, false friends → Week 08 drills · day 1 w08-comm-1

Agentic AI · 60 min · Eval tooling survey: promptfoo vs DeepEval vs Ragas vs Inspect — pick per use case (regression, RAG metrics, agent tasks) → Evals II: promptfoo, DeepEval, Ragas, Inspect in CI · resource w08-ai-1
DSA · 30 min · Diameter of Binary Tree + Balanced Binary Tree → Trees: DFS, BFS, BST · resource w08-dsa-2
Communication · 30 min · Vocabulary: Review and 100-word check → Week 08 drills · day 2 w08-comm-2

DSA · 25 min · Same Tree + Subtree of Another Tree → Trees: DFS, BFS, BST · resource w08-dsa-3
Architecture · 45 min · Balancing Coupling in Software Design: strength, distance, volatility — score couplings in the capstone → Coupling, cohesion & modularity (balanced coupling) w08-arch-1
Python · 20 min · Rep: class decorators vs metaclass vs init_subclass for a tool registry → Decorators, descriptors & metaclasses w08-py-2
Communication · 30 min · Speaking drill: Deliver a 5-minute presentation → Week 08 drills · day 3 w08-comm-3

Agentic AI · 60 min · LLM observability: Langfuse traces/spans/scores + OTel GenAI semantic conventions (still Development status) — instrument one request end to end → LLM observability: Langfuse, Phoenix, OTel GenAI semconv · resource w08-ai-2
Java/Spring AI · 30 min · Compare Spring AI vs Python RAG on the same 30-query set; note differences in retrieval defaults → RAG with Spring AI & pgvector w08-java-1
Communication · 30 min · Idioms & phrasal verbs: Idiom review and usage in speech → Week 08 drills · day 4 w08-comm-4

System Design · 45 min · Blob storage + CDN: presigned URLs, cache keys, invalidation, edge compute — for storing ingested documents → Blob storage, CDN & edge w08-sd-2
DSA · 25 min · Lowest Common Ancestor of a Binary Search Tree → Trees: DFS, BFS, BST · resource w08-dsa-4
Python · 20 min · Rep: PEP 695 generics (class Result[T]) + TypedDict for eval records → Advanced typing: generics, Protocols, ParamSpec, TypedDict · resource w08-py-3
Communication · 30 min · Writing: Writing an RFC introduction → Week 08 drills · day 5 w08-comm-5

Agentic AI · 120 min · Capstone build: GitHub Actions eval job (promptfoo or DeepEval) failing the build under thresholds; Ragas faithfulness/context-precision report on the golden set → Evals II: promptfoo, DeepEval, Ragas, Inspect in CI · resource w08-ai-3
DSA · 30 min · Timed coding mock (interview-2): Binary Tree Level Order Traversal + Binary Tree Right Side View in 30 min → Trees: DFS, BFS, BST · resource w08-dsa-5
Java/Spring AI · 30 min · Records + sealed types for eval results in Java; pattern-matching switch over outcomes → Modern Java 21→25: records, sealed types, patterns w08-java-2
Communication · 30 min · Record & shadow: Deliver a 5-minute presentation → Week 08 drills · day 6 w08-comm-6

System Design · 45 min · Timed 45-min AI SD mock (interview-2): design AI-powered semantic search for a 10M-document corpus → Design AI-powered semantic search · resource w08-sd-3
Architecture · 30 min · Write ADR 008 for capstone: pgvector over Qdrant for v1 (coupling, ops cost, scale ceiling, exit path) → Coupling, cohesion & modularity (balanced coupling) · resource w08-arch-2
Staff+ · 30 min · Write a stakeholder update on eval-in-CI: what it catches, cost, what you need from other teams → Communicating with executives & stakeholders w08-staff-1
Communication · 30 min · Soft skills + weekly review: Handling tough questions — checkpoint comm-2: re-record and compare against the baseline rubric → Week 08 drills · day 7 w08-comm-7
Review · 60 min · interview-2: full mock — SD + coding + AI SD + behavioral, score with the unified interview rubric and log gaps into next week's plan → Interview framework & back-of-envelope estimation w08-rev-1