Senior Software Engineer, Backend/infra

Pika Labs Pika Labs · AI Frontier · Palo Alto, CA · Engineering

Senior Backend & Infra Engineer to build and scale the core infrastructure for Pika's AI products, focusing on autonomous AI agents, their runtimes, orchestration, and integration with LLMs and multi-modal capabilities. The role involves architecting distributed systems, optimizing agent execution, designing real-time architecture, and building memory/retrieval systems.

What you'd actually do

  1. Architect Distributed Systems: Build scalable backend, infrastructure, and agentic services for Pika’s web, mobile, and multi-platform products.
  2. Evolve the Agent Runtime: Design and optimize the core execution loop that handles agent reasoning, tool-use frameworks, function calling, memory retrieval, and multi-step orchestration.
  3. Design Real-Time Architecture: Own and scale real-time messaging infrastructure, event-driven architectures, WebSocket connections, and pub/sub patterns for throughput, latency, and reliability.
  4. Implement Core AI Capabilities: Optimize LLM integrations, multi-provider model routing (Claude, GPT, Gemini, open-source), context window management, cost optimization, and streaming responses.
  5. Build Memory & Retrieval Systems: Design semantic search and vector-based embedding infrastructure to handle long-term memory, working memory, and episodic recall for autonomous agents.

Skills

Required

  • 5+ years of software engineering experience building production services at scale
  • 2+ years hands-on with LLM-based orchestration, multi-agent systems, or agentic solutions
  • Deep proficiency in modern backend technologies (Node.js, Python, Go) and frameworks (Express, FastAPI, TypeScript, etc.)
  • Strong understanding of distributed systems, event-driven microservices, message queues, cloud infrastructure (AWS/GCP), Kubernetes, and CI/CD workflows
  • Solid grasp of LLM capabilities and limitations, prompt engineering (system prompts, chain-of-thought, structured output), tool-use execution, and embedding models
  • Comfort designing and debugging real-time streaming pipelines, long-polling, and highly concurrent networking setups
  • Deep understanding of database design and the product sense needed to make an AI system feel "alive" and responsive
  • Ownership mentality

Nice to have

  • Experience with multi-modal AI architectures (image generation, TTS, speech-to-text, video generation)
  • Experience with agent frameworks (LangChain, CrewAI, AutoGPT) or building custom, high-performance execution runtimes
  • Experience with fine-tuning, RLHF, or DPO pipelines
  • Background in multi-tenant SaaS or internal tooling and operational automation
  • Previous startup experience—comfortable with ambiguity and rapid experimentation
  • Competitive coding background (IOI, ICPC, Olympiad medalists, etc.)

What the JD emphasized

  • 2+ years hands-on with LLM-based orchestration, multi-agent systems, or agentic solutions
  • Ownership mentality—identify systemic bottlenecks and ship solutions without waiting for exact specifications

Other signals

  • autonomous AI agents
  • cognitive agent runtimes
  • orchestration frameworks
  • LLM integrations
  • multi-modal AI architectures