AI Automation / Backend / Platform Engineer building DriveDesk as backend-owned AI automation and platform systems: RAG/LLM workflows, CRM/ERP integrations, FastAPI/PostgreSQL, Docker/CI, DevOps, runbooks, and recovery paths for real business processes.
Now open for remote-only full-time roles and fixed-scope projects where the output is a working system, not a demo shell. 10-second hiring signal: remote backend/platform engineer for AI workflow, CRM/ERP/API integration, internal tools, and DevOps reliability; fastest proof: DriveDesk AI Operator, DriveDesk Core, and AI Ops Workflow Kit; first useful result: one backend-owned workflow slice with state, tests, logs, docs, and handoff; team fit: English-first async review, international remote delivery, privacy boundaries, CI, runbooks, and operational handoff quality.
Message me first when: there is one messy workflow, one risky integration boundary, or one backend/platform/DevOps slice that should become testable, logged, documented, and handed off. First reply promise: I will answer with a fit read, risky assumptions, the smallest responsible first slice, the proof path to inspect, and the right next route: remote role, fixed-scope project, technical review, or no-fit.
Shortest inbound path: Message on LinkedIn -> Decision-Ready Contact -> LinkedIn Recruiter Packet -> Hiring Decision -> DriveDesk Proof Route -> AI Backend Proof Pack -> PDF resume. LinkedIn activation layer: LinkedIn Proof Signal includes a LinkedIn Activation Pack with copy-ready post variants, first-comment copy, short-DM reply copy, verified live skill surface, and the next skill-order target; Post B is selected as the role-fit/search post.
Recruiter shortcut / 30-second screen: message first through Decision-Ready Contact with role title, remote setup, stack, first-month ownership, timeline, and compensation band if shareable; then use the LinkedIn Recruiter Packet for inbound offer fit, shortlist signal, risk-to-proof routing, AI execution proof, copy-ready first messages, recruiter filter map, recruiter Boolean search strings, and message triggers, plus proof links before deeper checks through Hiring Decision, Recruiter Preferences for LinkedIn profile alignment and the search-intent proof stack, LinkedIn skills search surface, Role Targets, Skill Evidence, and the PDF resume. Best match: backend/platform, AI automation, CRM/ERP/API integration, internal tools, or DevOps ownership. International and Europe-compatible team fit: English-first docs, async proof routes, integration contracts, privacy boundaries, CI, runbooks, and operational handoff quality. Shortlist scorecard: real business infrastructure support, backend/platform implementation, AI workflow engineering, CRM/ERP/API integration, DevOps/recovery thinking, English-first async proof, and public verification discipline. Risk-to-proof screen: hidden automation state -> audit/outbox/retries; unverifiable AI output -> citations/eval/approval; brittle CRM handoff -> adapter contract/idempotency/dead-letter; remote delivery risk -> CI/smoke/runbooks/handoff notes. International role screen: distributed teams can inspect written proof, first-month ownership, privacy/integration boundaries, CI/live smoke, runbooks, and timezone-friendly handoff before a live technical call.
AI-native delivery loop: I use AI tooling to compress discovery, implementation, debugging, docs, and review into faster iteration, while the public proof stays engineering-owned: architecture, state boundaries, privacy checks, tests, logs, deployment, runbooks, and shipped quality remain my responsibility.
| Featured work sample | What it proves |
|---|---|
| DriveDesk Proof Route | Flagship backend-owned AI workflow: RAG, transcript analysis, approval, CRM handoff, audit, retries, and adapter boundaries. |
| AI Automation Role Fit | Precise fit for RAG, structured output, CRM/API handoff, tests, logs, and runbooks. |
| AI Backend Proof Pack | Compact technical screen across backend/platform, integrations, DevOps, verification, Decision-Ready Contact, and resume evidence. |
| Hiring Decision | Ownership signal: incomplete business context -> risky boundary -> small verifiable delivery plan with backend-owned state, AI-output review, CI, smoke checks, docs, runbooks, and handoff notes. |
| Enterprise Readiness | International and Europe-compatible async-team readiness: architecture, state ownership, integration contracts, privacy, CI, runbooks, verification, Decision-Ready Contact, and handoff quality. |
| Skill Evidence | Skill-to-proof map across Python, FastAPI, PostgreSQL, RAG, Docker/CI, CRM/ERP integration, DevOps, and AI-native delivery discipline. |
Employer trigger router: message me for AI workflow/RAG, CRM/ERP/API integration, backend/platform ownership, or DevOps reliability work when the first useful result should be a verified slice with tests, logs, docs, and handoff. Proof map: AI Ops Hiring Signal Brief, AI Ops Employer Trigger Proof, AI Ops First Slice Playbook, and Business Scenario Replay.
Search-fit titles: Back End Developer, Python Developer, AI Automation Engineer, Artificial Intelligence Engineer with a workflow/backend focus, Platform Engineer, DevOps Engineer, Integration Engineer.
LinkedIn recruiter search alignment: remote-only full-time roles where Python, FastAPI, PostgreSQL, Docker, GitHub Actions, RAG, Vector Databases, CRM/ERP/API integration, Systems Integration, DevOps, OpenAPI, n8n, Telegram, or pgvector appear in the role.
Recruiter filter map: title filters, skill filters, remote/location fit, language/review fit, featured proof, and seniority/specialist signal all point to public evidence instead of loose keywords.
Recruiter Boolean search strings: ("AI Automation Engineer" OR "LLM Workflow" OR RAG) AND (Python OR FastAPI) AND (CRM OR ERP OR integration), (Python OR FastAPI) AND (PostgreSQL OR pgvector) AND (Docker OR "GitHub Actions"), and ("Integration Engineer" OR "Systems Integration") AND (CRM OR ERP OR 1C OR banking OR webhooks).
Best immediate starts: AI workflow automation, CRM/ERP/API adapter, backend/platform slice, DevOps recovery sprint, or DriveDesk AI Operator-style proof route.
Fast public review: AI Backend Proof Pack, AI Automation Role Fit, DriveDesk Proof Route, Enterprise Readiness, Role Fit Pack, Flagship Platform, Selected Proof Projects for pinned proof repos, case studies, AI Ops Hiring Signal Brief, AI Ops Business Scenario Replay, AI Ops First Slice Playbook, AI Ops demo walkthrough, AI Ops live owner proof, and role fit. The 2-minute role-fit route is Enterprise Readiness -> Role Fit Pack -> Decision-Ready Contact -> PDF resume.
Inbound / contact: Message on LinkedIn with one role, workflow, or project; one success condition; and the systems involved. For recruiter offer triage, send role title, remote setup, stack/team surface, first-month ownership, hiring timeline, and compensation band if shareable. If you arrived from LinkedIn Contact info, use this portfolio as the landing route: recruiter hiring -> LinkedIn Recruiter Packet, project request -> Services, technical proof review -> AI Backend Proof Pack. LinkedIn Services requests route through LinkedIn Services Fit -> Decision-Ready Contact -> Inbound Brief before PDF resume. For low-friction outreach, Start Conversation and Decision-Ready Contact expose copy-ready first messages for remote role screens, workflow projects, and technical proof reviews. Decision routes: Decision-Ready Contact, Contact Routes, Inbound Brief, Work With Me, LinkedIn Services Fit, and Services.
Message reason router: use Start Conversation to choose remote role hiring, workflow/project, or technical proof review before sending LinkedIn context.
Fast fit checklist: message me when the work is remote-only and needs a backend-owned AI workflow, CRM/ERP/API integration, internal operations tool, DevOps recovery/hardening path, or one first working slice with tests, logs, docs, and a handoff route. I am not positioning for onsite-only roles, pure prompt/content tasks, isolated static websites, student/course assignments, standalone game clones, generic mobile/ecommerce apps without backend/integration ownership, or vague work where success cannot be defined.
| Send me | Best match | Fast proof |
|---|---|---|
| Remote full-time role | Back End Developer, Python Developer, AI Automation Engineer, Platform Engineer, DevOps Engineer, integrations, internal tools, Python/FastAPI/PostgreSQL, Docker/CI. | Hiring decision, Role targets, first month plan, Decision-Ready Contact, PDF resume |
| Fixed-scope project | One messy workflow, one success condition, and a working slice with tests, logs, docs, and handoff path. | Work with me, proof-backed fixed-scope offers, Start conversation |
| Technical proof review | Code, CI, live smoke, runtime boundaries, privacy redaction, adapter contracts, runbooks, and public-safe evidence. | Verification pack, AI Ops proof status |
DriveDesk AI Operator is the sharpest current proof route: documents, call audio, transcripts, or CRM leads -> RAG, analysis, scoring, tasks, Telegram approval, and CRM handoff with backend-owned state, audit, retries, idempotency, and adapter contracts. The larger DriveDesk Flagship Platform connects that slice to roles, records, workflows, adapters, audit/outbox, admin/operator surfaces, DevOps, observability, backups, and recovery paths.
Since March 2024, I have supported Autoschool54 digital infrastructure remotely while turning real operational problems into a broader platform path: admin workflows, documents, schedules, payments, accounts, access, automation, deployment, logs, backups, and recovery paths.
| Proof | What to inspect |
|---|---|
| Selected proof projects | DriveDesk AI Operator, DriveDesk Core, AI Ops Workflow Kit, DeployMate, and MPlusForm as inspectable project surfaces. |
| DriveDesk Core | FastAPI/PostgreSQL operations platform foundation: tenants, RBAC, audit/outbox, adapters, OpenAPI, Docker, pytest, CI, docs, public demo. |
| AI Ops Workflow Kit | RAG workflow backend: hiring signal brief, business scenario replay, document intake, first-slice playbook, committed demo GIF/walkthrough, privacy redaction before RAG/approval/CRM handoff, deterministic RAG quality eval (rag_eval=2/2), transcript analysis, approvals, Telegram callback, dry-run Bitrix CRM handoff, idempotent outbox, retry/dead-letter behavior, live PostgreSQL/pgvector runtime (storage=postgres), restart persistence proof, reviewer observability snapshot, 54-pass public verification gate, CI, and smoke checks. |
| AI Ops demo walkthrough | Short visual proof route for transcript -> RAG -> approval -> CRM-safe handoff, generated from the public-safe offer demo. |
| Live owner proof | Real Telegram approval evidence with CRM-safe outbox handoff and Bitrix dry-run boundary. |
| DeployMate | Self-hosted Docker deployment control panel: FastAPI/Next.js/PostgreSQL, SSH runtime tooling, CI/CD, health checks, evidence bundles, review packet, runbooks. |
| MPlusForm | Validation-boundary automation proof: Python sync pipeline, server-approved snapshots, Windows operations docs, Lua addon, and public verification run. |
| Need | My ownership surface |
|---|---|
| AI workflow / RAG automation | Ingestion, retrieval, analysis, scoring, approvals, operator docs, tests, and metrics. |
| CRM/ERP/accounting/banking integration | Explicit adapters, mappings, retries, logs, idempotency, rollout notes, and recovery path. |
| Internal ops platform | FastAPI/PostgreSQL backend, admin UI, roles, audit trail, workflows, deployment, runbook. |
| DevOps / self-hosting hardening | Docker, CI/release gates, health checks, logs, backups, smoke checks, and rollback procedure. |
LinkedIn: linkedin.com/in/alex-gerlitz-a659ab3bb · LinkedIn Services