Architect, Agentic Databases

At Teradata, we believe that people thrive when empowered with better information. Teradata Autonomous Knowledge Platform activates enterprise intelligence by unifying data, knowledge and business context to achieve tangible outcomes. With Teradata, organizations can provide agents with full context for impact when it matters. Our solution lets businesses connect and scale on premises, in the cloud, or through a hybrid approach. Teradata delivers real business value with AI.

Responsibilities

Own the data plane architecture — the Postgres engine, storage, query execution, WAL handling, extensions— and drive the Technical PoC that proves the design against live customer workloads; Define the end-to-end technical architecture of the Agentic Database data plane: Postgres engine configuration, storage model, WAL strategy, MVCC handling, extension framework, and the five Agent Services layered on top; Make foundational architectural calls on shared-storage vs. local-storage, separation of compute and storage, branching and copy-on-write semantics, and how Postgres extensions interact with the custom storage layer; Own the technical trade-off decisions — consistency vs. availability, performance vs. cost, scaling limits, durability guarantees — and document them clearly enough that engineering teams can execute without re-litigating them; Design the integration architecture between the Agentic Database and Teradata Fabric, Teradata Context Engine, and the Enterprise MCP/AgentStack platform, including CDC pipelines from the operational layer to Teradata’s OLAP engine; Own the Technical PoC end-to-end: design the test harness, select representative customer workloads, define success criteria, run benchmarks, and produce written findings that either validate the architecture or trigger a course correction; Serve as the technical authority in architecture reviews, design discussions, and cross-team dependency resolution; Review and approve technical designs from engineering leads; push back on architectures that trade long-term maintainability for short-term velocity; Mentor senior and staff engineers on database internals, distributed systems design, and production-quality engineering standards.

Requirements

A low-ego, experimental mindset who treats AI as a high-trust collaborator in their own work, navigates large complex codebases comfortably, and moves from architecture to working prototype without losing momentum; Deep, hands-on Postgres expertise: worked at the source level, understanding the query planner and optimizer (cost models, statistics, join ordering, partial indexes), the storage layer (heap, FSM, visibility map, TOAST), WAL internals, MVCC, vacuum, and the extension API; Experience navigating and modifying large C codebases; Systems programming proficiency in C or C++ for data plane work; familiarity with higher-level languages (Go, Python, Ruby) for control plane and tooling work; Strong understanding of storage engine design trade-offs: B-tree vs. LSM, page-based vs. log-structured storage, write amplification, read amplification, space amplification, and how each affects OLTP workload patterns; 12+ years of industry experience designing, building, and operating large-scale distributed systems, databases, or cloud data infrastructure; Demonstrated ability to own an architecture end-to-end: from initial design through technical PoC, production launch, and post-launch iteration with measurable outcomes at each stage; Experience building or evaluating high-availability, multi-tenant database systems in cloud environments, including automated failover, backup and restore, upgrade automation, and multi-tenant isolation; Comfortable functioning as both a problem solver and a problem finder; Strong written and verbal communication; BS/MS in Computer Science or related field, or equivalent depth demonstrated through production work; AI-native working style: use AI as a high-trust collaborator in engineering work, prototype faster, and look for ways to make systems more useful to AI agents; Practical understanding of agentic workload patterns: bursty parallelism, LLM-generated SQL, stateful session continuity, branch-on-demand isolation, and required database design decisions; Familiarity with the Model Context Protocol (MCP) and major agent frameworks (LangChain, LangGraph, OpenAI Agents SDK) at the level needed to define the integration surface between the Agentic Database and the agent execution layer.

Benefits

People-first culture; Flexible work model; Focus on well-being; Inclusive environment; Healthcare; Life and disability insurance; 401(k)-retirement savings plan; Time-off programs; Annual incentive plans