Forward Deployed Engineer

This is us
Kaltura’s (NYSE:KLTR) mission is to power any video experience for any organization – live, on-demand, or real-time. We not only want to make using video simpler, but we also want to better people’s lives through video. Founded in 2006, Kaltura is now a global leader in the video market with millions of people using our products daily to teach, learn, watch, connect, and collaborate. Among our customers, you’ll find more than 1000 global, well-known organizations.
15+ years since starting the company, we continue to foster a diverse and collaborative work environment where everyone gets a say. Our team is currently 700+ people, and we’re still growing. We have offices in New York, London, Singapore, and Tel Aviv, but our technology is all in the cloud.
Kaltura has a fast-paced environment where initiative is always encouraged. Together with our hybrid work model and flexible state of mind, you get the right conditions for creative juices to flow freely. Thanks to our long line of products, cultivation of rich collaborative culture and care for each Kalturian, you’ll never run out of room to grow and evolve.
If you don't meet 100% of the requirements below - that's okay, nobody's perfect! We believe in hiring people, not just a list of skills. We encourage you to apply if you think this is a role that would make you excited about coming to work every day.

Requirements:

The Role
You are the hands-on engineer in a Forward-Deployed Engineering pod — a small team embedded with priority customers, owning technical delivery from first prototype to stable production. You build customer-specific capabilities that deliver immediate value and become durable platform capability when proven broadly useful.
Your pod never modifies the platform core. You consume the Foundation team's shared components — registry, gateway, evaluation harness, guardrails, memory — and extend them for your customers. Everything you ship is a versioned, tested engineering artifact held to platform-grade quality. When a skill you built for one customer proves useful across others, it is promoted into the foundation.
Agent engineering is the discipline: prompt design, software engineering, and evaluation — combined. You write the behavioral instructions, write the code beneath them, and prove both against evaluation and latency cases. You do this while embedded with the customer, translating their domain into working, tested artifacts.

Scope
For each customer engagement, your pod delivers:

Skills — versioned artifacts: behavioral instructions, executable code (scripts, tool wrappers, data transformations), domain knowledge, tone, and jargon handling. Registered through the platform registry with mandatory evaluation gates.
Integrations — wrapping and registering customer tools, APIs, and MCP servers with contracts and permissions. Bringing each through the Foundation's gateway. Validating the full flow meets latency budgets — fixing and contributing back when it doesn't.
Evaluation suites — ground-truth scenarios and acceptance cases drawn from the customer's domain. No skill passes the registry gate without them.
Speech assets — ASR vocabulary biasing (so the system recognizes customer brand names, products, acronyms) and TTS pronunciation lexicons (so the avatar speaks them correctly).
Ongoing tuning — knowledge creation, behavior configuration, checkups, and production feedback integration as the customer's needs evolve.

What You Bring
Required

4+ years in a technical, customer-facing engineering role — Forward Deployed Engineer, Solutions Engineer with production code responsibility, or Software Engineer with consulting/deployment experience. You write and review production-quality code, not just configure.
Production experience with LLMs — prompt engineering, agent development (LangChain, LangGraph, or equivalent), RAG pipelines, tool/function calling, evaluation frameworks, and deployment at scale. You stay current with the latest capabilities and patterns.
Ability to author and evaluate AI skills as engineering artifacts — writing behavioral instructions, writing the code beneath them, designing test scenarios that catch real failures, and measuring behavioral consistency across runs.
Experience building integrations that started as customer-specific and became broadly reusable — wrapping external APIs, webhooks, and tools into governed, contracted interfaces. Understanding latency across distributed calls.
Strong domain learning ability — you absorb unfamiliar industries (finance, education, media, telecom) quickly enough to author domain-specific artifacts within weeks. You conduct discovery well and translate domain conversations into engineering requirements.
Clear communication with engineers and non-technical stakeholders alike — you serve as the senior technical counterpart for customer teams, conduct architecture reviews, and spot risks early enough to act on them.
Ability to scope work, sequence delivery, and remove blockers — making deliberate trade-offs between scope, speed, and quality to protect delivery timelines in fast-moving or ambiguous environments.

Strongly Preferred

Experience with MCP servers and agent frameworks — building or consuming them in production, understanding tool registration, contracts, and multi-agent coordination.
Background in speech/NLP pipelines — ASR vocabulary biasing, TTS pronunciation customization, or similar domain adaptation of speech models.
Experience with evaluation and observability tooling — DeepEval, Ragas, Langfuse, Opik, or similar. You've measured LLM quality beyond "does it look right."
Familiarity with real-time system constraints — understanding that a conversational avatar cannot stall, and how latency budgets shape what you can do within a turn.
Experience with multi-tenant platforms — tenant isolation, scoped configuration, registry patterns, and why customer A's skill must never leak into customer B's agent.
Track record of codifying patterns into tools, playbooks, runbooks, and reproducible benchmarks that scale a field engineering function beyond individual heroics.

What Success Looks Like

1 month: Domain immersion complete for your first customer. First skill authored, evaluation suite written, registry gate passed. Vocabulary biasing list delivered for ASR. Embedded relationship with customer team established.

3 months: Full skill set live in production for first customer. At least one customer tool integrated through the gateway with latency validated. Evaluation coverage sufficient for deployment confidence. Feeding learnings back to Foundation team on platform gaps.

6 months: Second customer onboarded independently (with pod support). Production feedback loop running — skills refined from real usage data. At least one skill promoted to the foundation as a reusable template. Patterns codified into playbooks for the next pod.

How Your Work Flows

You build (customer-specific) Registry gate (eval must pass) Production

Proven broadly useful Promoted to Foundation Requirements flow up via FDE Portfolio Owner
You never build your own authorization, failure handling, or evaluation infrastructure. You use the Foundation's shared components and contribute back when gaps appear. The boundary is what allows the platform to scale.

Why This Role

This is not consulting that ends with a handoff. It is not a support role that triages tickets. You are an engineer whose deliverables happen to be customer-specific — versioned, tested, production-grade, and promotable into the platform when they prove broadly useful.

Every deployment is an opportunity to make the platform better. Your work compounds.