AI Engineer (Back-end)

Taste is building the AI for creative work.

Today’s AI is great at logic and code, but when it comes to anything visual, brand, or aesthetic, it produces slop. Generic fonts, cluttered layouts, off-brand colors, soulless compositions. The models don’t understand taste.

We’re changing that. We’re building the creativity layer for AI: the models, evaluators, and data pipelines that teach AI what “great” actually looks like, so that every company building with AI can ship work that feels crafted, on-brand, and beautiful.

We recently closed a $16M seed round co-led by Amplify and CRV. We’re working with customers like Anthropic, Flint, and leading AI labs, powering the design intelligence behind their products. The team today is small, senior, and hungry. We care deeply about quality, move fast, and love the problem.

If you think AI-generated work feels soulless right now, and you want to be part of the team fixing that, we’d love to meet you.

We’re hiring a Backend-focused AI Engineer to own the infrastructure that powers our creative AI pipelines — scraping, indexing, embeddings, data platform, and the APIs that serve our models to customers.

You’ll work alongside our Applied ML Engineer and full-stack AI Engineer to make sure our data and serving layer can scale with the models we’re training. You’ll own architectural decisions on everything from our vector store to our training data pipeline.

WHAT YOU’LL WORK ON

• Data infrastructure: scraping, ingesting, embedding, and indexing massive visual + design datasets
• API architecture: internal services + external customer-facing APIs
• Training data pipelines: turning expert-labeled creative work into clean training datasets
• Model serving: inference infra, latency/cost optimization, reliability
• System design for an AI-first product that needs to scale fast

YOU MIGHT BE A FIT IF

• 4+ years backend engineering, recent experience at an AI-native startup or building AI infra in a product company
• Strong Python + typed TypeScript; comfortable with cloud infra (AWS/GCP)
• You’ve built high-throughput data pipelines (Kafka, Dagster, Airflow, whatever)
• Experience with vector stores, embeddings, RAG at production scale
• You think in systems and care about reliability

NOT A FIT IF

• You want to be in a pure-infra role with no product exposure
• You’re looking for a research-only environment
• You don’t enjoy collaborating closely with ML engineers on training data + eval

Compensation Range: $175K - $275K