Director / Sr. Engineering Manager - AI & Agentic Systems

Job description

**About Delhivery** Delhivery is India’s leading logistics and supply chain infrastructure company. We are a technology-first organization operating a massive physical network. Our proprietary platforms handle millions of shipments daily across 18,000+ pin codes, generating vast volumes of multi-modal, spatial, operational, and conversational data. As we transition toward autonomous, self-healing supply chain workflows, this role will lead the engineering frontier that builds our central intelligence and agentic platforms. **The Mandate** As the **lead for AI & Agentic Systems** , you will hold the architectural blueprint and operational authority for Delhivery’s AI ecosystem. Your mandate is to design the core platform infrastructure, establish engineering standards, and govern the deployment of autonomous agentic systems across all product lines (including Warehouse Management, Transportation, Last-Mile, and Merchant Platforms). You will act as the central technical authority, driving the adoption of open standards to securely and uniformly connect AI models to Delhivery's enterprise data sources and operational tools. You will enable product teams to build safe, deterministic, and highly optimized AI systems while reviewing and greenlighting all agents before they launch into production. **Core Responsibilities** - **Establish Enterprise AI Standards:** Own the definition of technical blueprints, design patterns, and SDKs for agentic workflows across the organization. Establish organization-wide standards for Agentic implementations, covering orchestration, context management, structured outputs, tool invocation, security, identity, observability, evaluation, lifecycle management, and interoperability. - **Centralized Architecture & Launch Review:** Serve as the final technical gateway for all AI agent deployments at Delhivery.Design automated governance frameworks that validate architecture compliance, tool permissions, security boundaries, model behavior, latency budgets, cost efficiency, reliability, structured output correctness, evaluation quality, and operational readiness before any AI system reaches production. - **Architect Production-Grade Agentic Workflows:** Design and deploy core multi-agent systems optimized for large-scale enterprise problems involving planning, reasoning, workflow orchestration, long-running execution, memory, human-in-the-loop approvals, asynchronous execution, and real-time decision making across warehouse, transportation, customer experience, and operational platforms. - **Build the Central AI Control Plane & Catalog:** Architect Delhivery’s shared AI infrastructure layer. Develop central enterprise capabilities for unified model routing, automated prompt drift detection, vector database multi-tenancy, and a **centralized catalog** that allows secure, audited discovery of enterprise data and services by autonomous agents. - **Lead a High-Caliber Engineering Team:** Mentor and scale a specialized engineering squad focused on foundational AI platform development. Cultivate a unified AI engineering community across Delhivery by hosting architectural reviews and upskilling product engineers on production-grade systems design and MCP-driven architectures. **Core Capabilities & Architectural Focus** This role is centered on building the underlying infrastructure, standardized protocols, and evaluation frameworks required to scale intelligent workflows safely: - **Platform & Gateways:** Designing resilient LLM gateways, automated model failover mechanisms, and enterprise-wide token optimization telemetry, inference optimization to deliver low-latency AI services. - **Evaluation & Deployment Pipelines:** Building continuous integration frameworks that programmatically test, benchmark, and audit the safety, retrieval accuracy, prompt regressions, safety, hallucination rates, policy compliance, and tool-calling reliability of agents developed across different business units before production deployment. - **State & Orchestration:** Engineering custom graph architectures or leveraging advanced state-management systems optimized for high-throughput, asynchronous tool execution. **What We Are Looking For** - **Proven Engineering & Governance Leadership:** 10+ years of core software engineering experience, with a significant track record of architecting, deploying, and establishing engineering standards for high-scale distributed systems or building production AI systems involving LLMs, agentic workflows, orchestration frameworks, evaluation systems, retrieval architectures, or model serving platforms. - **Platform Architecture & Protocol Expertise:** Demonstrated success in building developer platforms, shared infrastructure SDKs, or building systems leveraging open interoperability protocols (such as gRPC, GraphQL, or Model Context Protocol). - **Systems-First Pragmatism:** A deep commitment to engineering metrics—focusing heavily on system latency, API validation, compute cost-efficiency, secure tool execution boundaries, and predictable, deterministic software behavior. - **Cross-Functional Influence:** The ability to effectively collaborate with, advise, and align diverse product engineering heads and business stakeholders toward a unified technical vision.