The posting
【ABOUT US】
Raccoon AI is an emerging generative AI company. Our core product automates customer service messaging for e-commerce and online platforms, resolving 50 to 80% of conversations instantly.
The standard this role is accountable for is an autonomous agent that resolves real customer conversations end to end, at production quality and production volume, across languages.
【ABOUT THE ROLE】
This is a senior individual contributor role for an engineer who arrived at production work through research.
The foundation is a first-principles understanding of how these models actually work. When behavior breaks, that understanding is what turns guesswork into an explainable engineering decision, reasoning through tokenization, attention, sampling, context window, and training distribution. It is also what model strategy rests on: whether to stay on hosted models or invest in fine-tuning and self-training, a decision this role owns and defends with data, cost, and technical trade-offs.
What this role is hired for is that depth pointed at a product. Model selection and prompt tuning are the starting point, not the substance. The substance is the design of AI behavior for specific business contexts: defining what the agent is permitted to do in a given scenario, where its boundaries sit, when it must refuse or escalate, how it handles ambiguity and adversarial input, and how each of those decisions is measured. Scenario-level configuration of this kind determines whether the product is trusted in production.
The expected pattern of work is to read the literature, form a position, run the experiment, and land the result in production. Direction is set by this role rather than handed to it.
【WHAT YOU'LL DO】
- Own the behavioral design of the production AI agent: scope, permissions, guardrails, refusal and escalation policy, and failure handling for each business scenario
- Own LLM system architecture end to end, covering prompting, RAG retrieval, multi-agent tool use, and evaluation methodology
- Own model strategy, including hosted, fine-tuned, and self-trained options, argued from data and cost, with multi-vendor abstraction where it is warranted
- Debug model behavior from first principles: tokenization, attention, sampling, and context window effects
- Build quantifiable evaluation and observability so that every prompt or model change is measured before release
- Track frontier research and convert it into product decisions
- Raise the LLM engineering capability of the wider team
【WHAT YOU'LL NEED】
- A master's degree or above in Computer Science, Information Management, or an AI-related field, from a leading domestic or international university. A thesis in LLM or NLP is a strong signal. Equivalent depth demonstrated through production AI model work will also be considered.
- 5+ years in software or ML engineering, including 2 to 3 years building production LLM or GenAI systems
- Experience in both a large engineering organization and an early-stage startup. This role requires the engineering discipline of the former and the ownership of the latter.\
- Deep understanding of transformer architecture, attention mechanisms, tokenization, pretraining, fine-tuning, RLHF, and decoding strategies\
- A research foundation that has been carried into production. A thesis, publications, or a record of reproducing and extending papers, followed by shipped systems built on that understanding.
- Self-direction. This role defines its own problems and priorities rather than waiting for specification.
- The ability to explain model behavior and trade-offs to non-technical stakeholders and influence decisions with them
【NICE TO HAVE】
- Fine-tuning, training, or distillation experience (LoRA, SFT, RLHF/DPO)
- Multilingual and CJK NLP experience
- Inference optimization and inference serving
- Experience building LLM evaluation and observability systems
- Experience building or operating autonomous customer service agents (Intercom Fin, Sierra, Decagon, or equivalent)
