AI Data Software Engineer

Tiki AI is scaling the engine that powers the full model feedback loop for frontier AI development. This is an early-stage engineering hire at a company with real revenue and real frontier lab partnerships. We are looking for engineers who are technically strong, curious, and fast learners, and who want to grow quickly by building at the frontier of AI.

This is not a traditional software engineering role with a fixed backlog, nor is it a pure, siloed coding position. You will work as a hands-on engineer who helps turn frontier AI client requirements into real results. Alongside experienced engineers, you will learn to make sense of data needs that are often ambiguous, then follow the work all the way through: from understanding where a model falls short, to building the scalable, production-ready systems that address it, to showing that the model measurably improved. You will start where your skills fit best and take on more of the process over time, always working toward the same goal: real model improvement, not just delivered data.

WHAT YOU WILL WORK ON

• Requirement Deconstruction and Solution Design: Partner with top-tier AI research labs to translate frontier model training objectives, such as code generation, mathematical reasoning, and spatial understanding, into executable data production workflows and pipelines.
• Rapid Prototyping and Validation: Build standalone MVP demos, scripts, and workflows to independently collect, synthesize, or clean non-standard data, utilizing whatever tools or languages are necessary for swift validation.
• Scale and Operations Engineering: Package technical solutions into scalable SOPs for internal teams and vendors. Align technical frameworks with external LLM clients, collaborate on technical requirements, and help resolve workflow bottlenecks.
• Frontier Research and Evaluation: Stay ahead of AI research trends by digesting arXiv papers and technical reports. Direct algorithms like SFT, RLHF, and Chain-of-Thought into measurable, automated data production standards and agentic workflows.
• System Reliability and Architecture: Contribute to data pipeline design and improvements to increase system reliability, observability, and data pipeline throughput as programs scale., * A traditional ticket-taking SWE role: You will not be executing a fixed backlog or a highly managed product roadmap; you are responsible for bringing order to ambiguity.
• A remote-first position: In-person collaboration in San Francisco is vital to our speed and execution.
• A pure enterprise infrastructure scale job: Your focus is on agility, rapid closing of loops, and inventing data workflows for frontier AI, rather than polishing existing large-company infrastructure., This role is relevant to candidates currently holding or recently holding titles including: Software Engineer, Junior Software Engineer, Full-Stack Engineer, Backend Engineer, AI Engineer, Machine Learning Engineer (entry-level), Applied AI Developer, LLM Integration Engineer, Agentic Systems Developer, AI Product Engineer, Software Engineer (New Grad), API Engineer.

• Technical Background: Possesses a solid technical background and programming experience; capable of reading code, writing scripts, and assessing technical feasibility, without being dependent on a specific technology stack. Proficient in at least one programming language.
• Problem-Solving and High Agency: A proven ability to define the problem, find the right tools, and rapidly execute. You thrive when requirements are vague, resources are constrained, and you must proactively execute and deliver results in a fast-paced environment.
• AI/ML Domain Curiosity: Strong interest in AGI, LLMs, and agentic workflows. You should be eager to read academic papers and think critically about the underlying data architectures required to power modern AI.
• Location Requirement: Being a Bay Area local is required., * Hands-on experience with AI data production, synthetic data generation, or data engineering tailored for LLM training.
• Familiarity with alignment data paradigms like SFT, RLHF, or Chain-of-Thought, or data engineering for agentic reasoning models.
• Experience collaborating closely with product managers, technical account managers, or external partners.
• Bilingual proficiency: Mandarin.

Pulled from the full job description

• Parental leave
• 401(k)
• Health insurance
• 401(k) matching
• Paid time off
• Vision insurance
• Health savings account, Annualized base of $135,000 - $185,000

Full benefits including medical, dental, vision, and 401k are available., * 401(k)

• 401(k) matching
• Dental insurance
• Health insurance
• Health savings account
• Paid time off
• Parental leave
• Vision insurance

Tiki AI builds the infrastructure that teaches frontier AGI systems how to think, not just predict. We turn top-tier human expertise in STEM, code, and other domains into high-signal training data through cognitive extraction, going far beyond basic labeling. And we own the outcome, not just the delivery. We run the full model feedback loop: evaluate a model to find where its reasoning breaks, design data to fix it, prove the lift through training, then re-evaluate and go again. That is how we stabilize complex reasoning and fix long-horizon agent failures in production-scale systems, and how we report impact as a measurable improvement in the model. We lead with evaluation, because eval decides what to build, not just whether the last batch worked. That focus is driving our momentum: we are the trusted partner for several of the top-5 frontier labs, with exponential, multimillion dollar quarterly growth. We are hiring the team that will define how the world's most advanced AI models are trained and evaluated. If you want to work at the infrastructure layer of AGI, not the periphery, this is where that happens.