Forward-Deployed AI Engineer | Outcome Engineering | Mid-senior

You own an AI system from first commit to the client's reported outcome: built on their data, running inside their IT, measured against a number agreed before you wrote any code. Every enterprise has a demo that worked once. You build the one that is still working in month six.

Your Mission: The Role

Outcome Engineering is how Searce turns AI from a pilot into a line on the P&L.; As a Forward-Deployed AI Engineer, you own the whole chain: scope the outcome, prove it with working software, harden it for production, and stay until the number moves.

You will build with the best models from Google, Anthropic and OpenAI, route between them on merit, and be judged on outcomes, not demos. The agent is the last tool we reach for, not the first.

If you write production Python, can argue with a CFO about a number, and have opinions on eval harnesses you would defend in a code review, this is your seat. Bring the opinions. We bring real data and a problem that has a budget.

Your Responsibilities:

Prove it in days, not decks

- Own the POC end to end: Scope it against an outcome metric agreed upfront, build it on real client data, and demo working software.
- Win the room with code: Run workshops and demos for CxOs and engineers where the thing on screen is the thing being sold.

Build agents that survive production

- Architect the system: Orchestrate with Google ADK, Claude Agent SDK or OpenAI Agents SDK. Connect agents to enterprise data, APIs and business systems through standard tool and context interfaces (MCP). Handle sessions, memory, vector and state storage, and sandboxed tool execution.




- Make it deterministic where it must be: Output validation, tool routing, fallbacks and hybrid model routing.
- Ship the evals before the agent: Evaluation suites and tracing are part of the build, tracked as eval pass rate.

Ship it into the client's world

- Run it inside enterprise IT: Turn the reference architecture into a secure, observable production system. Success is tracked as POC-to-production conversion.
- Tune for performance and cost: Across Google, Anthropic and OpenAI stacks, tracked as cost per task.
- Stay until it is real: Monitor drift, bias and privacy, and report the client's outcome number back to the client.

Compound every deployment

- Be the engineer the CxO trusts: Say what not to build as readily as what to build.
- Leave patterns behind: Reference architectures, eval sets and agent templates that make the next Searce deployment faster. Track what is emerging across ADK, Claude Agent SDK and OpenAI Agents SDK, and adopt only what survives production.

Functional Skills

- Problem Framing: Turns a vague ambition into a testable hypothesis with a number attached.
- Systems Judgment: Knows when the answer is an agent, a workflow, a script, or "don't build it."
- Technical Storytelling:



Explains an agent's behavior and trade-offs to a CFO and to a staff engineer without changing the substance.
- Commercial Sense: Ties architecture choices to cost, revenue and risk.
- Ownership Under Ambiguity: Moves forward in messy client environments with no spec and no hand-holding.

Tech Superpowers

- Agentic Systems Depth: Hands-on with Google ADK, Claude Agent SDK, OpenAI Agents SDK or LangGraph. Fluent in MCP and A2A, session and memory management, and sandboxed execution. Computer-use agents are a plus.
- Evals and Observability: You have built evaluation harnesses and agent tracing for systems in production.
- Retrieval and State: RAG, vector stores, prompt caching, hybrid mode routing.
- Production Engineering: Strong Python, APIs and microservices. Delivery on GCP (Vertex AI), AWS (Bedrock, SageMaker) or Azure (OpenAI/Foundry), with enterprise security and responsible-AI practice (monitoring, bias, privacy).
- AI-Native Build Workflow: Ships daily with Claude Code, Cursor or GitHub Copilot, and knows where not to trust them.

Experience & Relevance

- 5+ years in software engineering, including 2+ shipping LLM or agent systems to production.
- Bachelor's degree in Computer Science, Engineering or a related field.
- At least one production agent, copilot or AI workflow with real users, and the metric it moved.
- Delivery in a complex or regulated enterprise environment.

Join the ‘real solvers’

ready to futurify?

If you are excited by the possibilities of what an AI-native engineering-led, modern tech consultancy can do to futurify businesses, apply here and experience the ‘Art of the possible’. Don’t Just Send a Resume. Send a Statement.

📌 Forward deployed AI engineer (Bengaluru)
🏢 Searce
📍 Bengaluru