Agentic Delivery Systems Engineer

At Aspen View, we are passionate about transforming the way organizations approach technology. We specialize in creating high-performing, nearshore IT teams to help North American clients innovate faster and more efficiently.

As we continue to grow, we’re looking for exceptional people to join our team and help drive impactful change across industries.

Aspen View is hiring an Agentic Delivery Systems Engineer to own how well agentic software delivery actually works on a large AI-led modernization programme for a global payments and financial-technology platform. The client is deploying AI-assisted development across 8,000+ engineers, running an agentic AI operating system for banking built on OpenAI with Amazon Bedrock and AWS Agent Core, and operating autonomous coding agents against core-platform modernization in parallel.

Its restaurant, lodging and hospitality commerce business is a named growth vertical.

Aspen View delivery teams build software with agents in the loop, and this role owns the part that decides whether that works: the context that reaches the model, the token and cost envelope it runs in, the tools and permissions each agent is given, and the evidence that any of it is making teams faster without making the code worse. This is a delivery system, not a chatbot — you will instrument first, form a hypothesis, and let the data overrule intuition.

• Design what enters the context window and what does not: retrieval strategy, file and repository scoping, structured summaries, and pruning rules that keep long sessions coherent across a large, heterogeneous estate.
• Diagnose and fix context degradation — window saturation, stale or contradictory context, lost-in-the-middle recall loss, and compaction that silently drops critical state.
• Optimise prompt caching, session compaction, and reuse to cut redundant token spend without reducing task success rates.
• Build and maintain cost models for agentic work: cost per task, per merged pull request, and per delivery sprint, with forecasts leadership can plan against under an active cost-reduction programme.
• Right-size model selection per task tier across frontier and smaller reasoning models — routing routine work down and reserving frontier capacity for genuinely hard reasoning.

• Own system prompts, agent instructions, and repository-level configuration as versioned, reviewed artifacts — not ad hoc text pasted between engineers.
• Design the tool surface each agent gets: tool inventories, MCP servers, custom skills, subagent decomposition, and permission and sandboxing models.
• Build internal MCP servers and integrations connecting agents to the systems delivery teams actually use — repositories, Azure Dev Ops Boards and Repos, CI pipelines, documentation, and data platforms.
• Integrate with the client’s existing AI estate rather than beside it:
  OpenAI APIs and the Codex toolchain, Git Hub Copilot, Microsoft Foundry, and agent runtimes on Amazon Bedrock / AWS Agent Core.
• Maintain a golden reference configuration set with a clear versioning, rollout, and rollback process across accounts.

• Stand up evaluation harnesses with golden task sets drawn from real delivery work — including payment, POS, and hospitality commerce flows — so configuration changes are proven rather than assumed.
• Run controlled comparisons across models, tools, and prompt strategies, accounting for run-to-run variance before declaring a winner.
• Instrument agent runs end to end: traces, token and latency telemetry, tool-call success rates, and cost attribution by team, account, and…