Танилцуулга

Mindset & AI Foundation

AI/LLM-ийн суурь ойлголт, промт, хамтын ажиллагааны mindset.

AI/LLM гэж юу вэ? — LLM-ийн дотоод бүтээгдэхүүн, хэрхэн ажилладаг вэ?
Промт гэж юу вэ? & Промт инженерчлэл — One-shot vs few-shot, context windows, system prompts.
AI-тай хамт ажиллах — Айдасаа өөрчлөх замаар давуу тал үүсгэх.
AI-ийг хөгжүүлэх — Өөрийн датагаар LLM-ийг цэнэглэх.
SDD & Vibe Coding — Specification-driven development; specs as source of truth.

Developer Tooling Upgrades

Хөгжүүлэгчийн хэрэгсэл, workflow, орчны сайжруулалт.

IDE & Editor — Cursor, VS Code + Copilot/Claude/Codex, AI-assisted editors.
Git & Github/Gitlab/Bitbucket — Best practices — Branching strategies, meaningful commits, PR culture.
Local dev environment — Parity between dev/prod without touching production DB.
Аюулгүй байдал ба автоматжуулалт — Secrets, linting, dependencies.

Testing Culture

Тестийн соёл, пирамид, TDD болон AI-тай тест бичих.

Яагаад тест хэрэгтэй вэ? — Why fear of AI breaking business logic is correct; how tests solve it.
Types testing — Frameworks, structure, naming; accounting/financial logic.
Testing pyramid ба TDD — Unit vs integration vs E2E; writing tests before AI-generated code.
Дэвшилтэт аргууд — Property-based, mocking, mutation, coverage, snapshot.
Testing with AI — Prompting AI to generate test cases you hadn't thought of.

CI/CD from Scratch

CI/CD-ийн үндэс, pipeline, deploy стратеги ба аюулгүй байдал.

CI/CD гэж юу вэ, яагаад хэрэгтэй вэ? — "It works on my machine" problem.
Pipeline үндэс — Build, test, deploy; first pipeline.
Environment promotion ба deploy стратеги — Dev → staging → prod; rollbacks; security.

Software Design & Techniques & Best Practices

Нийтлэг хэрэглэгддэг архитектур дизайний ойлголтууд

Event Driven Architecture — Queues, async processing, DLQ.
Concurrency ба lock-ууд — Idempotency, optimistic/pessimistic locking.
API design — REST conventions, versioning, idempotency.
Microservices vs Monolith — Microservices vs Monolith, when to use which.
Data Science — Data science & Data Warehouse & Data Pipeline & Tools.

Observability & Operations

Logs/metrics/traces, alerting/SLO, incident response.

Logs, metrics, traces — Three pillars; structured logging; APM.
Alerting ба SLO — What wakes someone at 3am; health checks; error budgets.
Incident response — Runbooks, postmortems, blameless culture.

Team & Process

Team-ийн процесс: code review, бичиг баримт, on-call, DoD.

Code review ба Git workflow — Giving and receiving feedback; trunk-based vs Gitflow.
Documentation ба technical debt — ADRs; technical debt as financial metaphor.
On-call, pair/mob, Definition of Done — Runbooks, pairing during AI sessions, formalizing done.

Хуучин кодоо яах вэ?

Одоо хүртэл ажиллаж байгаа чухал системүүдээ хэрхэн орчин үеийн болгох эсвэл хэзээ халах вэ?

Fire slowly by moving things to new services — Strangler fig pattern; extract to new services; incremental migration.
Start upgrading within itself — In-place refactoring; modernize inside the monolith first.
When to rewrite, risk, and testing — Rewrite vs. refactor; risk and rollout; testing legacy before touching.