LLM Solutions Architect

Job description

**ABOUT YOU** We are looking for an LLM Solutions Architect who is a builder at heart — someone who shapes strategy and ships real systems — to join our Monetization Products team. The best candidate will be someone who thrives in a fast-paced, highly collaborative, and exceptionally dynamic setting and is excited to drive AI product hypotheses from prototype to production-grade engineering. Strong technical architecture skills are essential, along with experience in designing and deploying LLM-powered systems in production. The ability to influence product direction, prototype rapidly, and communicate trade-offs clearly to both engineers and executives will be key to your success in this role. If you’re passionate about advancing AI technology solutions and love building intelligent, agent-first capabilities that transform how game developers monetize their products, we would love to hear from you! **ABOUT US** Xsolla is a global commerce company with robust tools and services to help developers solve the inherent challenges of the video game industry. From indie to AAA, companies partner with Xsolla to help them fund, distribute, market, and monetize their games. Grounded in the belief in the future of video games, Xsolla is resolute in the mission to bring opportunities together, and continually make new resources available to creators. Headquartered and incorporated in Los Angeles, California, Xsolla operates as the merchant of record and has helped over 1,500+ game developers to reach more players and grow their businesses around the world. With more paths to profits and ways to win, developers have all the things needed to enjoy the game. ### Responsibilities - Design end-to-end agentic architectures — tool-use schemas, intent parsing, multi-step orchestration, and safety guardrails — engineered for long-term ownership by product engineering teams, not solo maintenance. - Define the multi-modal interface strategy across our product portfolio: how the same capability is exposed via UI, API, SDK, and agentic natural language — consistently and without duplication. - Design the horizontal LLM platform layer — shared RAG pipelines, prompt libraries, vector search infrastructure, and evaluation frameworks — that product engineering teams can build on and operate independently. - Prototype rapidly to validate AI product hypotheses before full engineering investment. Prototype acceptance by product teams is a primary success signal. - Ensure every system you architect comes with the observability, documentation, and engineering runbooks needed for a product squad to take ownership confidently. - Shape product strategy alongside Product leadership: actively influence what AI capabilities get prioritized, in what order, and with what trade-offs. - Select and govern LLM providers and deployment strategies per use case — balancing cost, latency, accuracy, and privacy requirements. - Drive alignment across Engineering, Product, and Design on what ‘agent-ready’ means for each product surface. - Mentor engineers on LLM integration patterns, agent evaluation, and production deployment practices — building the team’s capability to own what you design. ### Qualifications & Skills **Required** - 5+ years of engineering experience, with at least 2 years designing and deploying LLM-powered systems in production. - Proven track record designing agentic systems: tool-use, function calling, multi-step reasoning, orchestration, and error recovery at production scale. - Experience designing AI systems for engineering team ownership — including observability standards, handoff documentation, and runbooks that let other teams maintain what you build. - Hands-on experience with major LLM APIs (OpenAI, Anthropic, Google Gemini) and at least one open-source model stack. - Experience building RAG pipelines with vector databases and orchestration frameworks (LangChain, LlamaIndex, or custom). - Strong Python engineering skills — production-grade LLM services, not just notebooks. - Demonstrated ability to influence product direction: you have shaped what gets built, not just how. - Clear communication in both directions: architectural trade-offs to engineers, business outcomes to executives. **Nice to Have** - Background in gaming, payments, or e-commerce — understanding of developer workflows, monetization models, or merchant operations. - Fine-tuning experience (PEFT/LoRA) for domain-specific model adaptation. - Experience with multi-agent orchestration frameworks (AutoGen, CrewAI, or custom). - Familiarity with LLM evaluation frameworks (RAGAS, DeepEval, or custom harnesses). - Exposure to EU AI Act, GDPR, or other AI compliance frameworks.