Agentic Engineering Lead (L3/L4)

ABOUT THE ROLE

We are hiring engineering leaders who deliver enterprise modernization and AI solutions with agents as the primary workforce and people as architects, reviewers and client partners. OptiSol is a Claude partner and works with AWS and Microsoft Azure. You will lead squads (formed per client engagement) and chapters (skill communities of 10-20 engineers) working on our iBeam (agentic application and data modernization), elsAi Foundry (agent governance and orchestration) and client solutions with agents embedded in them.

This is a full-time role in the Build & Modernize and Data & AI business units. The level is L3 (Squad Lead / Chapter Lead) or L4 (BU Lead / Engineering Manager), decided by your scope and leadership depth.

LOCATION

Chennai, Madurai or Coimbatore (Hybrid)

EXPERIENCE

L3: 5+ years. L4: 8+ years. This is indicative only, and enterprise depth matters more than years.

WHAT SUCCESS LOOKS LIKE IN YOUR FIRST YEAR

- Your squads deliver enterprise modernization with agents doing the bulk of the build, under explicit governance and human review.
- Engineers on your team grow into proficient Agent Engineers and are Claude-certified.
- Client CIOs and architects trust you with the solution architecture, not only the delivery.

KEY RESPONSIBILITIES Run delivery on an agentic SDLC

- Lead your squad so that BA, Dev and QA agents do the build work while people own intent, specification, review and acceptance.
- Own the instructions, skills and guardrails the agents run on, and the human-in-the-loop review queue that keeps their output trustworthy.
- Follow and improve OptiSol's Agentic SDLC SOPs, including the iBeam SOPs for application and data modernization.
- Track outcomes that matter: delivery against commitments, defects and rework, and cost per outcome including token spend.

Architect solutions with agents in them

- Design solutions where agents are used to build the system and where agents run inside it.
- Define how the agentic layer is built, deployed, monitored and governed in the client's own cloud or on-prem environment.
- Choose models, SDKs and cloud combinations per client constraint, and defend the trade-offs.

Lead enterprise engagements

- Work with client CIOs, enterprise architects, security teams and business owners from discovery through go-live.
- Run architecture reviews, solution walkthroughs and proofs of concept, and support proposals and estimates that reflect agent-assisted productivity.
- Spot modernization and AI opportunities inside accounts and bring them to business leadership.

Build the bench





- Coach developers to become proficient Agent Engineers and field-deployed engineers (FDE), and support their Claude certification.
- Hire, assess and grow engineers using hands-on, practical evaluations rather than self-assessment.
- Turn what your squad learns into reusable skills, patterns and accelerators for iBeam and elsAi Foundry.

WHAT WE ARE LOOKING FOR

- Enterprise delivery depth: 5+ years (L3) or 8+ years (L4) of software, data or AI engineering, with delivery on enterprise projects involving legacy estates, integrations, compliance and scale, and real stakeholder engagement with client IT, business owners, security and procurement.
- Claude-first, hands-on: you build with Claude Code and the Claude Agent SDK as your normal way of working, and can show real work such as specifications, skills, instructions, agents and evaluations you wrote. GitHub Copilot fluency is welcome.
- Clear thinker and communicator: you turn ambiguous business problems into precise specifications and explain trade-offs to a CIO in plain words.
- Deep in one area, credible across several: depth in application engineering, data engineering or AI/ML, plus enough breadth to guide teams outside your home stack. We do not screen on a checklist of technologies.
- Leader of people: you have led squads or teams, mentored engineers, and held a delivery commitment with a client.
- Cost and governance instincts: you treat token cost, security, auditability and human review as design inputs, not afterthoughts.
- Legacy modernization exposure (strong advantage): experience modernizing Oracle Forms, PowerBuilder, .NET, Java, or Oracle and Sybase to Postgres.

TECHNICAL AREAS YOU SHOULD BE COMFORTABLE WITH

- Agents in the build: run an end-to-end agentic SDLC with specification, BA, Dev and QA agents, human review and CI/CD. Engineer context, skills and instructions, and measure agent quality with evaluations.
- Agents in the solution: decide where an agent belongs and where deterministic code is better. Design tools and MCP servers, workflows versus autonomous agents, and state and memory.
- The agentic layer: build, deploy and operate orchestration, observability, versioned instructions and skills, identity and permissions, audit trails,



cost controls and human-in-the-loop gates.
- Agent SDKs: integrate the Claude Agent SDK, OpenAI Agents SDK, Google ADK and Microsoft agent frameworks, and choose between them. Use open protocols (MCP, agent-to-agent) so no vendor SDK becomes a lock-in.
- Models and clouds: combine Anthropic Claude (direct API, AWS Bedrock, Google Vertex AI, Microsoft Foundry), Azure OpenAI and Gemini per client estate, weighing data residency, private networking, IAM, quotas, latency and cost.
- On-prem and open-weight models: design for clients that need sovereign or on-prem models, benchmark open-weight models against a frontier baseline, and route work between them.
- Client-environment deployment: run the platform inside the client's cloud, VPC or on-prem network, with the client's repositories, ticketing and CI/CD, and with security and compliance sign-off.
- Model-agnostic design: keep models swappable behind a gateway so cost, capability and policy changes do not force a redesign.

L3 VS L4 - HOW THE LEVEL IS DECIDED L3 (Squad Lead / Chapter Lead): You lead one squad on an engagement or one chapter of 10-20 engineers. You run the agentic SDLC hands-on, tune agents and skills, and own squad outcomes. You design and defend the solution architecture for an engagement, act as technical lead with client architects and delivery owners, and coach developers to become proficient Agent Engineers. You contribute to pre-sales, proposals, PoCs and agent orchestration.

L4 (BU Lead / Engineering Manager): You lead a business unit, or several squads, chapters or a client portfolio. You set the delivery model and standards across squads and own governance and cost per outcome. You set reference architectures, model and cloud strategy, and approve high-risk designs.

You act as a trusted advisor to the client CIO and leadership and shape the account and the deal. You build and manage leads, plan capacity and succession, and run hiring. You own margin and utilization and lead business unit strategy, offerings and growth.

NICE TO HAVE

- Claude or GitHub Copilot certification
- Experience on an agentic modernization platform, or building agents for BA, Dev or QA work
- Hands-on modernization of Oracle Forms, PowerBuilder, or database migrations such as Oracle or Sybase to Postgres
- Experience serving or fine-tuning open-weight models on-prem or in a client's private cloud
- MLOps, DataOps and data governance experience
- Cloud certifications on AWS or Azure
- Pre-sales, RFP and estimation experience; talks, articles or open-source work on agentic engineering

📌 Agentic Engineering Lead/ (Chennai)
🏢 Optisol Business Solutions
📍 Chennai