Enterprise Architect

Job Summary

As an Agentic Forward Deployed Engineer (FDE), you will operate at the critical intersection of deep full-stack engineering, high-level system architecture, and client-facing consulting. You are not just writing code; you are bridging the gap between technical pre-sales and scalable, enterprise-grade production delivery. Embedded directly with enterprise clients, you will act as a primary technical advisor, turning ambiguous, messy business realities into production-ready autonomous and multi-agent systems—collectively known as Business Transformation Agents.

At the 8–12 years experience level, you will own the end-to-end system design and architecture of these solutions. You will navigate complex client environments to wrangler disparate data, design robust ETL pipelines, rapidly prototype Proof of Concepts (POCs), and carry them through to highly secure, integrated production ecosystems. Beyond core delivery, you will also play a pivotal role in pursuits, collaborating with account teams on discovery workshops, technical demos, and solution shaping, while leading and mentoring a lean engineering pod.

Key Responsibilities

• Pre-Sales to Production Bridging: Own the transition from rapid conceptual prototyping and discovery POCs to production-grade, highly scalable enterprise deployments, balancing velocity with operational readiness.
• System Design & Architecture Ownership: Systems Thinking, Define, model, and document the end-to-end architectural views (logical, data flow, integration, security, and operations) of agentic systems within complex enterprise ecosystems.
• Data Wrangling & ETL Integration: Architect and build ingestion, data modeling, and data pipelines to clean, structure, and orchestrate messy client data from disjointed legacy systems before applying AI/LLM orchestration.
• Build End-to-End Agentic Solutions: Apply hands-on engineering to design and deploy custom agents. Implement core agentic patterns: prompt/context engineering, Retrieval-Augmented Generation (RAG) or context graphs, memory, tool/function calling, Model Context Protocol (MCP), and multi-agent orchestration.
• Enterprise-Grade Integration: Connect AI agents to existing enterprise legacy databases, ERPs, and cloud environments using secure APIs, OAuth, custom enterprise connectors, and robust identity access management (IAM) controls.
• Evaluation & Testing Rigor: Establish rigorous testing suites, regression pipelines, and evaluation harnesses to continuously check agent safety, response quality, latency, and operational guardrails.
• Own AgentOps & Production Telemetry: Build CI/CD, deployment versioning, and real-time observability feedback loops to manage cost, trace errors, and monitor production agent behavior.
• Technical Leadership & Pod Governance: Set the engineering direction, conduct rigorous code/architecture reviews, mitigate delivery risks, and mentor a lean pod of 3–5 developers to build high-performance, reusable IP.
• Business Consulting: Prototyping (POC-to-Prod): Transitioning early-stage ideas and pre-sales requirements into secure, high-value, demo-ready POCs. Business interface and driving requirements / ideation with Business/ Running stakeholder discovery sessions, mapping workflows, building business cases. Eventually, Design and deploy Business Transformation Agents with measurable ROI. Also, drive efficiency through reduced cycle times, lower manual effort, and higher accuracy.

Skill Requirements

Other Requirements

• Fluency across multiple agent frameworks and architectural judgment to select the ideal model/tooling configuration based on specific client constraints.

• Hands-on experience deploying AI agents onto managed enterprise runtimes (e.g., Vertex AI Agent Engine, Bedrock AgentCore) with active cost-containment and governance patterns.
• Deep functional domain knowledge in core transformation spaces: Finance operations, procurement, supply chain, HR, insurance claims, customer service, or regulatory compliance.
• A proven track record of creating reusable code components, utility libraries, or specialized accelerators adopted across wider engineering communities.
• Broader AI/ML Python ecosystem familiarity (e.g., data prep, feature extraction, evaluation metrics) and comfort collaborating directly with data science specialists.

How Success is Measured

POC-to-Production Conversion: The efficiency and reliability with which pilot projects/POCs are successfully architecturalized, productionized, and integrated into live environments.

Sustained Value Creation: Quantifiable improvements in process cycle-time, manual effort reduction, increased accuracy, and decreased cost-to-serve for the client.

Architectural Stability & Security: Zero major security breaches or infrastructure regressions, combined with low exception rates, robust cost-management, and high system availability.

Team & Asset Velocity: Successful enablement and mentorship of the local delivery pod, alongside high utilization of reusable internal code accelerators.