The Role The Agentic Delivery Manager is accountable for how work gets from an approved request into production: planned, estimated, released, measured and kept inside its funds. You'll run delivery for a squad of product engineers building a real-time customer decisioning platform for a leading UK telecoms group. The platform decides which offer, message or next step each customer sees, across three consumer brands and every channel.

Part of the team is software. You run delivery through AI agents as well as people: the agents you set up do the chasing, reporting and routine checks, and you spend your time on judgement.

This is a client-facing consulting role requiring confident engagement with senior client stakeholders, sound judgement under ambiguity, and the credibility to challenge scope, spend and quality.

What the role owns. Delivery planning, estimates and resourcing, release coordination, spend tracking, delivery KPIs, risks and dependencies, and the agent-assisted delivery workflows.

The Business

Consultant shapes the vision and roadmap with the client and owns problem framing and acceptance.

The Lead Technical

Consultant owns architecture, engineering decisions and technical risk.

The Site Reliability

Engineer leads technical incident response and Product Support owns case handling, while the Delivery Manager tracks incidents and defects and makes sure the lessons reach the team.

What we expect of everyone. Every role on this team, consultant or engineer, works with a product engineering mindset and an AI engineering mindset. You own the outcome, not the ticket: you understand the business problem, shape the requirement, make and record the decision, ship in small steps and check the result in production.

AI is part of how you work, not a side project: you use AI assistants and agents every day, build with them where they help, check what they produce and share what works.

What makes it different

- Outcomes over ceremony. SAFe and Scrum are tools, not rituals. You keep what helps work flow and drop the rest.
- People and agents. You plan, assign, check and measure the work of AI agents alongside the engineers.
- You own the numbers. Every plan is modelled against the funds, the business principles and the team's real productivity, not just the ask.
- You understand what ships.



You know the domain well enough to challenge a requirement, trade scope or stop a release.
- You automate yourself. In your spare time, you hand the next part of your own job to a tool.

Key Accountabilities

- Own delivery for the squad from approved request to production, including plans, sequencing and dependencies.
- Give quick, defensible estimates and resource needs for recent demand, and re-plan the same day when things change.
- Run release management: release plans, environment readiness, go or no-go, cutover, rollback, release notes and post-release checks.
- Connect the delivery toolchain (Jira, Confluence, GitLab, CI/CD and incident tools) to AI assistants and agents, for example Claude through MCP connectors.
- Hand agents the routine work, such as status reports, ticket hygiene, triage, release notes, dependency chasing and meeting follow-ups.
- Set the rules for agent work: what an agent does alone, what a human reviews, and how its output is checked and measured.
- Maintain a live view that links demand, investment, capacity, delivery, risk and outcomes, so leaders can query it instead of trawling Jira and Confluence.
- Track spend against the allocated budget every week, forecast to the end of each funding period and flag overspend risk early, with options.
- Model every plan against the funds, the business principles and the team's measured productivity before committing to it, and say "not yet" when the ask doesn't fit.
- Measure the squad against a small set of agreed KPIs, such as throughput, cycle time, release success and right first time.
- Track right first time at feature level: how often a feature is touched again after it first reaches production, whatever the cause.
- Follow incidents and defects through to root cause with the SRE and Product Support, and feed the lessons back to the engineers and into the backlog.
- Learn the business domain in depth (offers, eligibility, pricing, ranking,



campaigns and channels), and make judgement calls on scope and sequencing with the Business Consultant: know when good enough ships and when it doesn't.
- Report delivery, spend, KPIs and risks to senior client and company stakeholders, briefly and early.
- Lead product engineers in the UK, nearshore and offshore: clear blockers the same day and spread good AI practice across squads.

Strong candidates may come from roles such as:

- 8-10 years of experience as Engineering managers or technical leads who have taken on delivery, budgets and stakeholder reporting.
- Platform, DevOps or product operations leads who have automated delivery workflows.
- Delivery or programme managers from SaaS, digital product or platform businesses who build their own automation.
- Consultancy delivery leads who have run fixed-price or fixed-capacity engagements against KPIs.
- Service delivery managers of large distributed teams, where they can show software product delivery and automation.

Useful Experience

- Customer decisioning, personalisation, next-best-offer, CRM, pricing or marketing platforms, for example Pega Customer Decision Hub.
- Telco, retail or consumer financial services.
- Cloud and event-driven platforms: AWS, GCP, Kafka and microservices.
- Running nearshore and offshore teams, for example in Romania and India.
- Building agents or copilots for a delivery team, or rolling out AI tools at team scale.
- FinOps, or cloud and AI cost management.
- Certifications are welcome but not required: we weigh evidence over badges.

What success looks like

First 90 days

- Your squad's status, spend and KPIs report themselves from live tools, with no hand-built decks.
- Routine delivery tasks run through agents or automation, each with a named human check.
- A release rhythm the client can plan around, with right first time measured from your first release.
- A weekly forecast of spend against budget to the end of the funding period.

First six months

- Right first time and escaped defects improving against your 90-day baseline.
- No surprise overspend: every variance flagged ahead, with options.
- You're asked for a view on scope and sequencing, not just dates.
- At least one of your automations reused by another squad.

📌 AI Technical Consultant (Bengaluru)
🏢 Moolya
📍 Bengaluru