AI Deployment Engineering Manager (all genders)

Job description

Who We Are Accenture Song is the world’s largest tech-powered creative group. Within Song, our Design & Digital Products practice helps iconic brands reinvent the products, services, and spaces that define their relationship with customers — from defining product strategy and crafting human-centred digital experiences, to engineering the AI-powered platforms and design systems that bring them to life at scale. We operate like a studio: small teams, high craft standards, and a direct line between design decisions and shipped products. Our team spans Product Design, Product Management, Engineering, Experience Design, and Content Design across 70 Song Studios worldwide. AI is not a feature layer we add — it is the architecture we build around. Who You Are You will embed directly into Accenture's priority accounts across Central and Eastern Europe. You are the technical owner on the ground, scoping problems, building solutions, closing the AI last mile, and demonstrating what AI-native delivery actually looks like. You are a senior technical practitioners who owns outcomes end-to-end and pulls AI tools forward into real delivery contexts and impact. You work at the intersection of engineering, product thinking, and client delivery. The expectation is that you can go from a vague client problem to a working proof of concept in days - and from there to a production-grade deployment in weeks. You have deep roots in software engineering and a genuine instinct for design — you understand that in Song’s studio model, the distance between an engineering decision and a user experience is deliberately short. You bring that awareness to every line of code and every client conversation. The Work Accenture Song is a radical commitment to the outcome. You will embed directly inside client organisations — sitting with their product teams, attending their standups, learning their systems, and building alongside their engineers. You will carry your laptop into rooms where decisions get made and your code into production environments where consequences are real. You will move from problem framing to working prototype to production deployment within a single engagement — leaving behind systems that run, documentation that teams can maintain, and patterns that the practice can scale. You will then move to the next client and do it again. Build & Deploy in Client Environments Production-First Engineering: Design, build, and ship production-grade AI applications directly inside client systems — including custom agents, LLM-powered workflows, MCP servers, sub-agents, and agent skills that integrate with the client’s existing technology estate. Full-Stack Delivery: Own the complete delivery stack: backend logic, API integrations, front-end surfaces, and cloud infrastructure. Whatever blocks the work becomes your job to write — there is no hand-off queue. Rapid Prototype to Production: Compress the gap between concept and working software. Run short, sharp prototype cycles to validate AI approaches, then drive them into production without losing the rigour that enterprise environments demand. Systems Integration: Connect AI-native capabilities to the client’s existing platforms — integrating APIs, data pipelines, identity systems, and enterprise tooling to ensure new AI workflows operate within, not alongside, the client’s reality. Embed, Operate & Own Deep Client Embedding: Operate inside the client’s environment as a trusted technical partner — attending team ceremonies, earning access to production systems, understanding organisational dynamics, and building the relationships that make complex deployments succeed. Diagnosis Under Ambiguity: Scope messy, under-defined problems into tractable engineering challenges. Identify where AI-native approaches will cre ate genuine leverage — and where they will not. Production Ownership: Stay accountable for what you ship. Monitor, debug, and optimise deployed systems; handle incidents when things break; and ensure stability before you hand over to the client’s own team. Human-AI Experience Engineering: Translate complex AI behaviours — reasoning traces, confidence states, agentic actions, fallback flows — into transparent, controllable user experiences. Work in tight collaboration with Song’s Experience Designers to close the gap between model output and usable product. Advise, Translate & Scale Technical Storytelling: Communicate technical trade-offs and deployment realities clearly to both engineering teams and executive stakeholders. Make complex AI architecture legible without dumbing it down. Pattern Extraction: Identify what is reusable. Document the architectures, integration approaches, and deployment patterns that worked — and feed them back into Song’s D&DP engineering practice as accelerators for the next engagement. Client Capability Building: Leave clients stronger than you found them. Run code-with sessions and workshops that transfer knowledge, raise the technical bar of the client team, and ensure the systems you built can be maintained after you leave. Practice Intelligence: Bring the field back to the studio. Surface what clients are actually asking for, what is breaking in production, and what AI capabilities are ready for enterprise deployment — informing Song’s offering development and go-to-market thinking. Extensive experience in engineering delivering production software — either as a software engineer with significant client-facing or consulting exposure, or in a prior forward deployed engineering role. Demonstrable track record of shipping AI-native systems into production: LLM-powered applications, agentic workflows, RAG pipelines, or multi-agent orchestration using frameworks such as LangChain, LangGraph, LlamaIndex, CrewAI, or AutoGen. Full-stack engineering capability — strong backend skills in Python (primary) and at least one of TypeScript or Java; comfort building front-end surfaces when the engagement requires it. Experience deploying to cloud environments (AWS, GCP, or Azure) via CI/CD pipelines, infrastructure-as-code (Terraform, Helm), and containerisation (Kubernetes, Docker); ability to own production operations end to end. Strong data engineering fundamentals — pipelines, structured and unstructured datasets, reasoning about data quality in messy real-world environments Experience designing and deploying enterprise AI applications including agentic workflows, RAG pipelines, vector databases, embeddings, and semantic search Demonstrated ability to scope ambiguous, real-world problems, move quickly from diagnosis to working software, and operate with high autonomy inside unfamiliar client environments. Strong cross-functional communication — able to run design workshops, present to technical and executive audiences, write clear documentation, and build trusted relationships with client engineering teams. Fluency in English — able to embed with DACH clients, facilitate workshops, and represent Accenture Song at the highest level in client environments. Bonus Points Experience building and delivering technical artefacts specifically associated with the FDE model: MCP servers, sub-agents, agent skills, custom connectors, or production LLM evaluation harnesses. Background in a prior forward deployed, solutions engineering, or technical consulting role at a product company, AI lab, or professional services firm — with a pattern of multi-week client embeddings and end-to-end delivery ownership. Design literacy: familiarity with design systems, component libraries, or human-centred design processes; comfort collaborating with Product Designers and Experience Designers as an equal creative partner. Experience with responsible AI in production — guardrails, content filtering, human-in-the-loop design, auditability, and bias evaluation — applied to enterprise-grade AI systems. A portfolio that demonstrates the FDE mentality: shipped systems, not slide decks. Open-source tools, p