Full Stack AI Engineer

What We’re Hiring For This is a full-stack platform engineering role with a strong automation and execution-system specialization. The engineer should be strongest in the systems that allow SynthralOS to actually perform work: Workflow execution Connectors and integrations Web scraping Browser automation RPA-style execution Background workers Data acquisition API/backend systems Full-stack product integration Infrastructure supporting these capabilities You do not need to be the primary architect of our advanced AI systems. However, because these systems interact closely with agents, retrieval, AI-assisted workflows, and AI-generated actions, you should have enough AI application engineering knowledge to understand how those systems behave and how your work connects to them. Core Working Speciality These are the areas where we expect strong hands-on production experience. Workflow & Automation Systems You should be comfortable working on: Visual workflow builders DAG/workflow execution Workflow compilation Node execution State persistence Long-running workflows Wait/resume behavior Human approvals Retries Replay Idempotency Branching Loops Conditional execution Scheduled execution Failure recovery Execution tracing Workflow debugging Background processing Workspace-level concurrency You should understand the difference between building a workflow editor and building the runtime that reliably executes what the editor represents. Connectors & Integrations Strong practical experience with integration systems is important. This includes: REST APIs OAuth Webhooks API keys Connector manifests Dynamic action schemas Trigger schemas Pagination Rate limits Credential handling Error normalization Connector versioning Connector health Generated operation catalogs Integration testing We prefer engineers who think in terms of connector classes and reusable patterns, rather than writing isolated fixes for individual integrations. Web Scraping & Data Acquisition You should be comfortable building and operating production web-acquisition systems. Relevant experience includes: Static HTML extraction JavaScript-rendered websites Crawling Structured extraction Content normalization Pagination Site-specific variability Data-quality validation Extraction retries Domain behavior profiling Persisted scraping jobs Routing between acquisition strategies Failure classification Confidence-based escalation You should understand that reliable scraping is primarily a routing, normalization, observability, and failure-recovery problem, not just an HTTP request problem. Browser Automation Production browser automation experience is highly relevant. You may work with systems involving: Playwright Puppeteer Remote browsers Browser workers Persistent browser sessions Page navigation Dynamic interactions Form completion Authentication flows File uploads File downloads Screenshots Data extraction Browser queues Browser retries Session recovery You should understand when a task can be solved using normal HTTP/data acquisition and when it needs to escalate into a real browser session. RPA & Action Execution Experience building systems that perform multi-step actions across external applications is particularly valuable. Relevant concepts include: Action sequencing Stateful execution Browser actions API actions Human-in-the-loop steps Approval checkpoints Credentials and permissions Execution recovery Retry boundaries Action verification Side-effect safety Long-running tasks Execution audit trails The goal is reliable computer and application interaction, not brittle macro scripting. Backend & Distributed Execution You should be highly comfortable with: Python FastAPI Async Python Pydantic SQLModel / SQLAlchemy PostgreSQL Supabase Redis Job queues Background workers WebSockets or streaming Scheduled jobs Event-driven execution Distributed state Concurrency controls Database migrations API design You should be able to trace a production task across: API → Database → Queue → Worker → External System → Persistence → Product UI Full-Stack Product Engineering This is still a full-stack role. You should be capable of working across: React TypeScript Vite TanStack Router TanStack Query Clerk Tailwind CSS shadcn/ui Graph/workflow interfaces Dynamic forms Node configuration panels Streaming execution interfaces Debugging surfaces Run histories Approval interfaces We do not expect every candidate to be a frontend design specialist. We do expect you to be capable of taking a platform capability from backend implementation through the actual user experience. Educational / Working Knowledge These areas are important for understanding the broader SynthralOS architecture, but they do not need to be your deepest speciality. AI Application Engineering You should understand the fundamentals of modern AI application systems, including: LLM application architecture Tool calling Structured outputs Prompt boundaries Agent execution Multi-step reasoning systems Model routing Context windows Token budgets Retrieval-augmented generation Embeddings Knowledge retrieval Agent memory concepts Human approval around AI actions AI evaluation Hallucination and failure modes Prompt injection risks Tool-result grounding Cost and latency tradeoffs You do not need to be an AI researcher. You should understand these systems well enough to safely integrate workflows, connectors, browser automation, data acquisition, and external actions with them. Agent Runtime Concepts Working knowledge of the following is useful: Agent orchestration Agent/tool loops Execution checkpoints Pause/resume Cancel and steer Conversation history Context assembly Evidence-backed outputs Approval gates Durable execution events Knowledge grounding Quality evaluation Agent actions inside workflows Advanced development of these systems may be owned by other engineers, but your work will frequently intersect with them. RAG & Knowledge Systems You should understand the basic architecture behind: Document ingestion Chunking Embeddings Vector retrieval Metadata filtering Knowledge indexing Retrieval pipelines Document provenance OCR-to-knowledge pipelines Deep specialization is not required, but you should understand how the data systems you build eventually become usable context for AI and workflow execution. Good to Have These capabilities are useful but are not the primary reason we are hiring this role. Traditional Full-Stack / CRUD Application Engineering Experience building conventional SaaS applications is useful. Examples include: CRUD interfaces Admin panels Dashboards Forms Tables User management Settings systems Billing interfaces Search and filtering File management Role-based access REST-backed application screens This experience provides a useful product-engineering foundation, but SynthralOS involves substantially more runtime, execution, integration, and distributed-system complexity than a conventional CRUD application. Fast AI-Led Development, Deployment & Bug Fixing We value engineers who know how to use modern AI-assisted development tools effectively without surrendering engineering judgment. Useful experience includes: Rapid repository exploration AI-assisted debugging Generating implementation drafts Refactoring with AI assistance Writing tests quickly Investigating logs with AI support Generating migration or integration scaffolding Comparing implementation against PRDs Rapidly tracing unfamiliar systems Accelerating repetitive engineering work Producing deployment and debugging checklists Using AI to reduce time-to-fix during production incidents The expectation is AI-assisted engineering, not AI-dependent engineering. You should still be capable of: Reading the code yourself Understanding the system architecture Validating generated changes Testing your work Diagnosing production failures Rejecting incorrect AI suggestions Understanding the blast radius of a change Speed is valuable only when paired with correctness. Strong Bonus Experience Additional bonus points for engineers who have worked on: Automation platforms Integration platforms Low-code builders RPA systems Workflow engines Browser automation products Developer tools Data acquisition platforms ETL/data pipeline systems Internal application builders Multi-tenant SaaS infrastructure Distributed task execution Document processing systems Ideal Skill Profile The ideal candidate might roughly look like this: Deep / Working Speciality Workflow systems Connectors & integrations Scraping Browser automation RPA Python backend systems Distributed workers PostgreSQL / Supabase Full-stack platform implementation Production debugging Strong Working Knowledge React / TypeScript Authentication and permissions Docker / Linux Deployment infrastructure Job queues Document extraction Security boundaries Multi-tenancy Educational / Architectural Understanding AI agents LLM tool use RAG Embeddings Model routing AI evaluations AI quality controls Agent orchestration Bonus Traditional CRUD SaaS development Fast AI-assisted coding AI-assisted debugging Rapid deployment workflows Infrastructure troubleshooting What We Do Not Need We are not specifically looking for: A pure frontend engineer A CRUD-only full-stack developer A pure ML researcher A prompt engineer Someone who has only built simple LLM wrappers Someone who only knows how to connect APIs together Someone whose automation experience consists primarily of scripting one-off tasks We need someone who understands how to turn automation capabilities into reliable, observable, reusable platform infrastructure. Role Summary Your speciality is execution. Workflows need to run. Connectors need to perform real actions. Scrapers need to acquire reliable data. Browsers need to complete tasks. Workers need to recover from failures. Approvals need to resume correctly. The frontend needs to accurately expose what the backend is doing. And all of those systems need to coexist safely with an increasingly AI-driven product architecture.