Member of Technical Staff (Software Engineer, Connector Platform)

Perplexity AI2 days ago

Location

San Francisco

Type

Full Time

Salary

USD 220,000 – 405,000

Level

Senior

Role

Backend Engineer

Posted

Jun 15, 2026

Full TimeSenior

The role

Summary

The Connector Platform team at Perplexity AI is seeking a Member of Technical Staff (Software Engineer) to design and build the critical data layer that enables AI agents to reliably access and execute integrations across hundreds of heterogeneous systems. This role requires deep expertise in backend systems architecture, API platform design, and authentication/authorization patterns to create a unified, typed interface for agent-driven enterprise integrations. You'll own end-to-end responsibility for connector runtime, semantic layers, tool discovery, robustness patterns, and quality frameworks that make agents more effective by providing grounded, permissioned access to real-time data and enterprise systems.

What you'll do

Connector Runtime Architecture and Implementation: Own the complete design, implementation, and evolution of the connector runtime system that registers, hosts, and executes diverse connector types (native, MCP servers, CLI-backed tools) behind a unified agent-facing interface. Build robust registration mechanisms, execution orchestration, and lifecycle management while maintaining type safety and backward compatibility.

Semantic Layer and Knowledge Management: Architect and expand the semantic layer including tool schemas, entity schemas, capability metadata, and relationship modeling. Implement mechanisms for capturing and applying organization- and account-specific corrections and knowledge, ensuring the platform becomes the source of truth for institutional knowledge rather than fragmenting it across individual connectors.

Tool Discovery and Selection Optimization: Design and implement the tool-discovery and tool-selection surfaces that agents use to identify and invoke appropriate connectors. Optimize for both model accuracy in tool selection and context efficiency, enabling agents to make intelligent decisions about which tools to call and when, directly improving agent reasoning quality.

Agent Loop Robustness and Reliability: Implement comprehensive robustness patterns including structured result formatting, partial-failure semantics, intelligent retry logic, idempotency guarantees, pagination handling, rate-limit management, and deep observability into every agent-initiated tool call. Ensure agents can handle gracefully degrade and recover from integration failures.

Authentication and Authorization Framework: Define and implement comprehensive authentication, authorization, and credential-isolation patterns for connectors including OAuth flows, BYOK (Bring Your Own Key) support, and per-org/per-account credential boundaries. Partner closely with Security and Backend Platform teams on defense-in-depth strategies to protect sensitive credentials and enforce proper access controls.

Connector Onboarding and Quality Assurance: Build the end-to-end connector onboarding pipeline including schema specifications, test fixtures, and comprehensive evaluation suites. Establish measurable quality standards and metrics that validate connector functionality within actual agent loops, replacing hope-driven deployment with data-driven quality gates and continuous validation.

Reliability, Observability, and Incident Response: Set and maintain the technical bar for connector reliability and operability by defining SLAs, implementing comprehensive monitoring and observability, establishing error-rate baselines, and building incident response playbooks. Ensure the high-fan-out integration surface remains always-on, performant, and predictable for mission-critical agent operations.

Cross-functional Platform Strategy and Collaboration: Partner with product, AI, and security teams to define clear connector interfaces, establish integration patterns, and drive platform evolution. Collaborate to ensure new agent capabilities can reliably build on the shared platform while maintaining backwards compatibility and consistent developer experience.

What we look for

Technical

Backend Systems Design and ImplementationDemonstrated experience designing and building production-grade backend systems at scale. Strong track record of architecting efficient, reliable, and scalable systems that handle operational complexity, with particular strength in platform-style systems serving heterogeneous downstreams.

API Integration and Platform ArchitectureHands-on expertise building or working extensively with API gateways, integration platforms, or connector systems that abstract multiple heterogeneous backends. Experience with managing API versioning, backward compatibility, and evolving interfaces without breaking clients.

Production Backend LanguagesStrong proficiency in at least one statically-typed or performance-critical backend language (Python, Go, or Rust). Ability to work effectively in multi-language environments and make pragmatic language selection decisions based on technical requirements and team capabilities.

Cloud Infrastructure and KubernetesHands-on experience with modern cloud infrastructure platforms such as AWS, Google Cloud, or Azure. Practical knowledge of containerization, Kubernetes orchestration, service mesh patterns, and operational best practices for running distributed systems in production.

Security Patterns and Protocol ImplementationIn-depth expertise in at least one of: OAuth 2.0/OIDC protocol implementation, authorization frameworks, credential management systems, or secure secret storage. Comfort working in security-sensitive areas and making informed trade-offs between safety guarantees, implementation simplicity, and development velocity.

Education

Computer Science FoundationBachelor's degree in Computer Science, Computer Engineering, Software Engineering, or equivalent professional experience demonstrating deep foundational knowledge of algorithms, data structures, and systems design principles.

Distributed Systems KnowledgeStrong theoretical and practical understanding of distributed systems concepts including consistency models, eventual consistency, idempotency, failure modes, and recovery strategies.

Experience

Backend Systems EngineeringMinimum 4+ years of professional experience designing and building backend systems deployed in production environments. Mid-level candidates typically have 4-6 years; senior engineers and staff engineers should demonstrate proportionally deeper impact and architectural leadership.

API and Connector IntegrationDemonstrated experience building, maintaining, or significantly contributing to API integration systems, connector frameworks, or MCP (Model Context Protocol) server implementations. Familiarity with the operational and semantic challenges of connecting diverse third-party systems.

LLM-Based Agent Tooling and EvaluationExperience building tooling, evaluation frameworks, or supporting infrastructure for large language model-based agents. Familiarity with how agents discover, reason about, and invoke tools, and the systems-level considerations that make agent loops robust and effective.

Distributed Systems and Operational ExcellenceTrack record of building systems that operate reliably at scale with attention to observability, error handling, rate limiting, pagination, and graceful degradation. Experience setting and maintaining SLAs for complex production services.

Skills

Required skills

Backend System ArchitectureDesign and implement scalable, reliable backend systems with deep understanding of trade-offs between consistency, availability, performance, and operational complexity.

API Design and Integration PatternsExpert-level API design encompassing schema evolution, backward compatibility, versioning strategies, and creating intuitive interfaces for heterogeneous consumer needs.

Python, Go, or RustProfessional-level proficiency in at least one statically-typed or performance-critical backend language, with ability to write maintainable, well-tested, performant code.

Cloud Infrastructure OperationsHands-on proficiency with AWS, Kubernetes, containerization, infrastructure-as-code, monitoring, and operational best practices for distributed systems.

Authentication and AuthorizationIn-depth understanding of OAuth 2.0, OpenID Connect, JWT, API key management, role-based access control (RBAC), and credential isolation patterns in multi-tenant systems.

System Observability and DebuggingExpert ability to implement comprehensive logging, metrics, tracing, and alerting to enable rapid debugging of complex distributed systems in production.

Pragmatic Problem-SolvingAbility to navigate ambiguity, balance competing engineering concerns (safety, simplicity, velocity), and make informed trade-off decisions alongside experienced team members.

Nice to have

MCP (Model Context Protocol) ExperienceHands-on experience developing, deploying, or managing MCP servers or understanding the protocol and its integration patterns with AI systems.

LLM Agent DevelopmentExperience building or contributing to frameworks, tools, or evaluation systems for large language model-based agents, function calling, or tool use patterns.

Third-Party API Integration at ScaleProduction experience integrating and managing many heterogeneous third-party APIs, handling API versioning, rate limits, and maintaining reliability across diverse upstream services.

Schema and Semantic ModelingExperience designing flexible, evolvable schema systems, semantic models, or metadata frameworks that enable tooling and automation across complex domains.

Type System DesignFamiliarity with building well-typed abstractions, potentially including experience with language-level type systems or schema definition languages (TypeScript, Protocol Buffers, GraphQL).

Security-First EngineeringTrack record of building security-sensitive systems with attention to threat modeling, defense-in-depth principles, and secure secret management practices.

Evaluation Framework DevelopmentExperience designing and building evaluation, testing, or quality assurance frameworks that provide quantitative confidence in system behavior across diverse scenarios.

Compensation & benefits

Salary

USD 220,000 – 405,000 (annual)

Stock options

Available

Interview process

1
Initial Screening Call — Preliminary conversation with a recruiter to discuss background, motivation for the role, and alignment with the Connector Platform team's mission. Typically 30 minutes and serves to confirm basic qualifications and cultural fit.
2
Technical Architecture Interview — Deep-dive system design conversation with a senior engineer focused on evaluating your backend architecture skills, approach to designing scalable systems, and experience with API integration or platform problems. Expect discussion of past projects, trade-offs, and design philosophy.
3
Backend Implementation and Coding — Hands-on technical interview assessing your proficiency in your chosen backend language (Python, Go, or Rust). Typically includes building a small service, implementing integration patterns, or solving a design problem with code.
4
Security and Infrastructure Knowledge — Technical interview covering authentication patterns, authorization frameworks, credential management, cloud infrastructure, and operational considerations for production systems. May include discussing OAuth flows, secret management, or infrastructure-as-code approaches.
5
LLM and Agent Context Discussion — Conversation exploring your understanding of large language models, AI agent frameworks, tool use patterns, and how system architecture decisions impact agent effectiveness and reliability. No specific prior LLM experience required; willingness to learn is key.
6
Team and Cross-functional Collaboration — Interview with potential teammates or cross-functional partners from product, security, or AI teams assessing collaboration style, communication clarity, and ability to navigate ambiguous problems alongside experienced engineers.
7
Final Discussion with Leadership — Opportunity to discuss role expectations, career growth, impact potential, and answer questions with the engineering manager or tech lead overseeing the Connector Platform team.