Skill-to-evidence map
Skills mapped to public evidence.
This page turns the profile skills into reviewable evidence: AI automation, Python/FastAPI backends, PostgreSQL, RAG workflows, n8n/Telegram orchestration, CRM/ERP integrations, Docker/CI handoff, runbooks, and validation boundaries.
Fast reviewer path: pick the skill, open the evidence link, then inspect code, CI, docs, demos, smoke checks, or runtime evidence.
AI Automation
Python
FastAPI
PostgreSQL
RAG
n8n
Telegram
CRM/ERP
Docker
Docker/CI handoff
AI Automation / RAG
Document/transcript/lead intake, privacy redaction before RAG/approval/CRM handoff, RAG retrieval with deterministic quality eval, transcript analysis, lead scoring, approval queues, Telegram callback approval, CRM-safe handoff, outbox drain, and n8n integration boundaries.
Open AI Ops role requirements map or privacy boundary
Backend / Internal Tools
OpsDesk Reviewer Replay proves the first-role slice: FastAPI, SQLAlchemy, PostgreSQL migration, ticket intake, operator queue, status transitions, idempotent outbox, SLA worker, SQL metrics, pytest, Docker Compose, GitHub Actions, smoke check, and privacy audit.
Open OpsDesk Reviewer Replay or DriveDesk Core review
CRM / ERP Integration
Explicit adapter contracts, field mapping, validation, retries, idempotency, audit logs, dry-run CRM handoff, rollout notes, and recovery thinking.
Open fixed-scope offers
Docker/CI Handoff
Docker runtime work, CI/CD, release gates, health checks, logs, deployment control panel evidence, runbooks, and recovery paths.
Open DeployMate evidence snapshot
Validation Boundaries
Client-side data, approved snapshots, sync pipelines, operational docs, Windows automation, and separation of untrusted input from accepted state.
Open MPlusForm reviewer snapshot / Open public verification run
Remote Operations
Autoschool54 backend/application-support work since March 2024: support workflows, docs, backups, troubleshooting, deployment/recovery thinking, and AI-native delivery discipline.
Open role / project brief
AI-Native Delivery Discipline
AI tooling accelerates discovery, decomposition, implementation options, debugging, docs, tests, and review. The engineering responsibility stays with me: architecture, state, integration boundaries, privacy, verification, deployment, logs, runbooks, and shipped quality.
Fast Problem Decomposition
Unclear business requests become workflow maps, risky assumptions, first slices, and verification routes.
Decision-ready contact
Verification Habit
Outputs are checked through code inspection, tests, smoke routes, CI, live evidence, docs, and runbooks.
Verification pack
Full-Cycle Delivery
Problem shape becomes backend slice, integration boundary, deployment path, operator handoff, and next-phase plan.
First month plan