Senior AI Agent Engineer - Developer - SmartBotics Digital Network

Feladatok:

What You'll Actually Do

Design, build and deploy AI agents and multi-agent pipelines for B2B clients across sales, support, and operations

Architect integrations between LLM layers, voice AI (VoIP/SIP), CRMs, and business workflows

Lead technical scoping sessions with clients — translate vague business problems into precise specs before a single line of code is written

Maintain structured project documentation: system diagrams, decision logs, API specs, prompt versioning, test results

Participate in code and architecture reviews; hold the bar on quality for your own work and the team's

Identify reusable components and contribute to the internal SmartBotics toolchain

Support pre-sales with honest, accurate technical assessments — no overselling

Elvárások:

What We Require

Hard skillsProduction-deployed AI agents — not demos, not proofs-of-concept

LLM orchestration: LangChain, LangGraph, AutoGen, CrewAI, or comparable

Prompt engineering at a system level: structured outputs, function calling, multi-turn memory

Python (primary); familiarity with JS/TS for integrations

REST API and webhook architecture; OAuth flows

Vector stores (Pinecone, Weaviate, pgvector) and RAG pipelines

Voice AI or telephony integration (Twilio, Bland, Vapi, or similar) — strong plus

n8n, Make, or comparable automation platforms

Working styleYou document before you build — specs, diagrams, edge cases

You surface problems early, not after a sprint is wasted

You can hold a client conversation without a salesperson present

You treat scope creep as a technical risk, not a favour opportunity

You understand that a quietly running production system is the goal — not a demo that impressed someone

Transparent about progress, blockers, and timelines — in writing, regularly

Important: We will ask you to walk us through a live production system you built — how it works, how it was scoped, what broke, and what you'd do differently. No live system = no further process. This role is not suited to candidates whose portfolio lives in notebooks, YouTube tutorials, or agency slide decks.

Előnyt jelent:

Reference Requirement — Non-Negotiable

As part of your application, please provide at least two delivered, live references in the following format for each:A brief description of what was built and what business problem it solves

A working URL, demo link, or verifiable deployment — something we can actually access or call

A direct phone number of the client contact who commissioned the work (reachable, will be called)

Your specific technical contribution — do not list team work as solo work

Stack and key architectural decisions with brief rationale

Reference submissions that provide only agency names, NDAs, or 'available on request' will not progress. We respect confidentiality agreements and will work with you on format — but we need something real and verifiable.

Example of a qualifying reference submission

AI-powered inbound call handler for a dental clinic chain (Budapest, 12 locations). Handles appointment booking, FAQ, and escalation to human staff. Live since Q3 2024, handling ~350 calls/month. Built on Vapi + GPT-4o + Google Calendar API. Client contact: Dr. Kovács Péter, +36-30-XXX-XXXX.

Has a live number to call   *    Has a named, reachable client   *    States exact contribution

Amit kínálunk:

Compensation offer: Honestly? We don't know yet.

SmartBotics is a young, ambitious company — currently 6 people. We don't work from salary grids. What we're looking for is someone who fits the team in terms of personality, makes us meaningfully stronger technically, at the time of decision represents the best value for what we're building, and will still play an important role in our life, when we are a 30 people team. That's the only formula we run.

So: tell us what you'd expect to start cooperation.

In your application, please include: the engagement model you prefer (employee, contractor, freelancer), and what you consider realistic and motivating compensation. Fixed monthly? Hourly rate? Project fee? Revenue share? A combination? We're open to all of these in principle — just be specific. If we see the solution in you, we'll make it work.

Our office is in Budapest, District XI. We prefer meeting in person when possible, but we are open to remote arrangements with occasional on-site days in Budapest.

What you get from usDirect access to leadership — no bureaucratic layers between you and decisions

Real projects: live deployments, real clients, real stakes — not internal tools or demo environments

A team that treats preparation and documentation as craft, not overhead

Transparent feedback — you'll always know where you stand and why

Room to shape how engineering works as we scale — this is not a role where you execute someone else's spec

Variety: different industries, different AI stacks, different levels of complexity on every project

The satisfaction of building things that actually run in production and make a measurable difference for clients