Job overview:

At Ross Research Labs, we’re pushing the limits of AI for real-time systems — blending agentic intelligence with Model Context Protocol (MCP) to bridge automation, control, and creativity. We design systems that think, coordinate, and act across distributed environments — and we’re looking for a developer who’s passionate about turning these concepts into reality. You’ll design and create production-grade AI features end-to-end, with a strong focus on agentic systems and the Model Context Protocol (MCP). You’ll help us design and build MCP servers/tools that expose capabilities, context, and control surfaces to agents and applications. Expect a mix of deep systems work (C++/Go), ML/AI integration (Python), and robust service/API engineering (gRPC/REST/WebSockets/Webhooks/MQTT).

Who you report to: Director, Ross Research Labs/AI

What we offer:

Ross offers competitive salaries, comprehensive health plans, and several perks to help you perform at your best. Some of these perks include flexible hours, generous paid time off, fitness/wellness allowance, an employee share ownership program, development support, and a ton of fun social activities and events! Best of all, you will be part of the Ross Video family, and we've got an energizing environment here.

What the job is all about:

• Design & implement MCP components: author new MCP servers/tools, schemas, and adapters; define capabilities, resources, and prompts; harden auth, routing, and observability.
• Build agentic features: implement planners, tool-use flows, state management, retries/guardrails, and evaluation loops without tying us to any single framework.
• Create production services: develop reliable backends and microservices in C++, Go, and Python with clean interfaces and strong contracts.
• API engineering: design and consume gRPC and REST services; implement WebSockets, Webhooks, and MQTT for streaming and event-driven interactions.
• Data & context plumbing: integrate vector search, RAG, and structured knowledge to extend domain knowledge for generic LLMs; maintain efficient context budgets while optimizing latency and cost.
• Develop intelligent content classification workflows across video, audio, and text, leveraging LLMs, VLMs, and ML models with contextual metadata to enable semantic and contextual understanding, tagging, and retrieval.
• Performance & reliability: profiling, optimize inference and tool calls, add tracing/metrics, and design graceful degradation/failover.
• Security & safety: enforce security for MCP tools, and apply content/behavioral guardrails.
• Code quality & teamwork: write tests, reviews, and docs; mentor peers; keep the bar high for maintainability.

Who you are:

• Hands-on MCP experience: you’ve designed or extended MCP servers/tools (or similar plugin/tooling protocols); you understand capabilities, resources, prompts, sessions, and transport.
• Agentic systems expertise: strong grasp of patterns like planning, tool orchestration, reflection/verification, retries, and evaluation — no framework lock-in required.
• Languages: strong in C++, Python, and Go. • APIs & realtime: production experience with gRPC, REST, WebSockets, Webhooks, and MQTT.
• Systems mindset: comfort with concurrency, streaming I/O, memory/CPU profiling, and building services that actually scale.
• AI integration: practical experience wiring LLM/VLM capabilities into products (prompting, output schemas, function/tool calling, evaluation).
• DevOps basics: containers, CI/CD, environment isolation, secrets, and observability (logs/metrics/traces).
• Ownership: you design, iterate, measure, and improve.