AI Solution Engineer
# AI Solution Engineer **Born Digital** · Praha · `On-site` 🕒 **Статус:** *Опубликовано: 6 дней назад* · *Источник: Indeed* --- ### About the Role Prague / Bratislava / Remote Elsewhere this role is called a Forward Deployed Engineer. Banks, telcos and utilities across ten countries run millions of conversations through our AI agents. Millions more still go through their people — calls, emails, tickets, recordings nobody can read at scale. So we build both: the agents that handle the conversations, and the analysis layer that reads all of them — for intent, sentiment, compliance and root cause, then does something with what it finds. Escalate it, route it, trigger the next process, brief the next agent. Getting them there takes engineers who don’t stop at the API boundary. You’ll sit inside our most strategic accounts — in their systems, their data, their release cycle — and own the path from “interesting pilot” to “in production, handling live traffic”. Most of that is configuration on our platform. The rest is code you write: MCP servers that expose the customer’s systems to the agent, LLM analysis pipelines over unstructured data, and custom extensions where the platform doesn’t reach yet. Knowing which of the two a problem calls for is the core skill of this job. About you You think about the whole chain, not just your part of it — how data moves between systems, what breaks first under load, and where the delay is actually coming from You understand the product and its architecture fluently — what each layer does, where the extension points are, what a change costs. Enough Python and API literacy to prototype, debug production and review what AI tooling writes for you You’ve built or integrated against MCP servers, or you’ll be dangerous with them within a week — you know how to expose a system, and a dataset, to a model You can turn a vague business question into something an LLM can evaluate across ten thousand conversations — and you know how to check whether the output is actually right You know how to present information — dashboard, report, whatever the audience needs. AI will draw the chart; choosing what belongs on the page is still on you Comfortable in someone else’s environment: enterprise VPNs, on-prem deployments, legacy CRM/CC stacks, integrations that are documented badly or not at all You’ve shipped something customer-facing and stayed to keep it running Hands-on experience with LLMs and agent frameworks is a must, and so is analytical thinking. Voice experience — STT/TTS, real-time audio, telephony or contact centre platforms — is a strong plus Customer-facing credibility: the CIO trusts your judgement on cost and risk, their engineers trust that you actually know the stack. You do that in Czech/Slovak and in English Bias to ship. You’d rather have something running on Friday than a perfect design doc Two centres of gravity here. Some of you come from the agent and voice side — STT/TTS, real-time audio, telephony, contact centre platforms. Some come from the data side — pipelines, LLM analysis, evaluation, reporting. We hire both. Deep on one, genuinely curious about the other. Key Responsibilities ✔️ Own end-to-end delivery of flagship deployments — from first integration spike to production traffic and measurable business outcomes ✔️ Configure agents on our platform — flows, prompts, tools, orchestration, guardrails, handover to human — and tune them until the numbers are right ✔️ Build and run the MCP servers that connect the agent to the customer’s world: CRM, core banking, billing, contact centre, knowledge bases, internal APIs — and that expose their data back to us ✔️ Design and build the analysis layer: what we extract from unstructured conversations (sentiment, intent, compliance, root cause), where it lands, and what it triggers — notifications, escalations, the next process, the next agent. A finding that ends in a PDF without producing an action doesn’t count ✔️ Write custom extensions where the platform doesn’t reach — and know when not to ✔️ Debug live production issues across the stack: LLM, speech and voice channels, data ✔️ Optimise cost and latency — model choice, token spend, and knowing what needs an LLM and what needs a regex ✔️ Prototype fast in front of the customer; turn the ones that work into something we can support ✔️ Spot when the same custom tool shows up in a third deployment and make the case for it becoming a platform capability — you’re the sharpest input our product team has ✔️ Work daily with the customer’s engineers and their leadership, and with our AI Deployment Leads on what to build and why Why join us? ️ You’ll build things that millions of real people actually talk to — and you’ll hear about it when it works More interesting projects than people to take them: banks, telcos, utilities across ten countries Problems nobody has solved yet, which is the fun part and occasionally the annoying part ️ Internal hackathons — no theme, no slides, just build something and show it on Friday ️ Two weekend teambuildings a year, plus the smaller unofficial ones nobody organises but everyone attends Genuinely good merch, restocked regularly — our clients wear it too A young team that genuinely likes each other, argues about architecture and means none of it personally Remote and flexible, competitive pay, no drama about either We’re building scalable platforms for AI-first products. Work onsite or remotely, whatever helps you deliver your best work. Side by side AI Solution Engineer (FDE) AI Deployment Lead (Deployment Strategist) Account Lead (Client Partner) Owns That it works That it matters That it lasts Core question How do we build and run this? What should we build, and for whom? Where is this account going, and what is it worth? Primary output Configured agents, MCP servers, integrations, analysis pipelines, dashboards, custom extensions Use case and measurement definition, scope, success metrics, rollout and adoption plan Account strategy, stakeholder map, value case, proposals and pricing, expansion plan Spends the day with Customer engineers, our platform, production logs Business owners, ops leads, C-level, our Solution Engineers Decision-makers and C-level, our Deployment Leads, sales Depth Deep in the stack — architecture, data, voice, failure modes Deep in the workflow — process, volumes, cost, what’s worth measuring, org politics Deep in the account — business, org, politics, commercials, portfolio fit Often comes from Backend or integration engineering, data engineering, voice and contact centre platforms Consulting, business analysis, BI and analytics, contact centre or back-office operations Enterprise account management, client partnership, consulting, solution sales Measured by Time to production, reliability, containment, quality of what’s built Adoption, business outcome, expansion into the next use case Retention, recognised customer value, account growth, quality of relationships Commercial role None — supports scoping with effort estimates Shapes scope and pricing alongside sales Owns proposal, pricing and negotiation for new work Number of accounts 1–2 deep; more when the work is analysis rather than integration 2–3, embedded A small portfolio, held for years When it goes wrong It’s live but nobody trusts it, or it breaks under load It’s built beautifully and nobody uses it The project delivered and the account still churned — or we heard about the problem too late Reports into Delivery / Engineering Product Commercial
Наблюдалась 2026-09-15, впервые 2026-09-09, источник — Indeed.