Software Engineer - Applications

Description

The company is an early-stage AI research and product lab building frontier multimodal models. It was founded by former DeepMind researchers, including Andrew Dai, who previously led engineering on Gemini.

Its first model launched publicly in October 2026.

The team has 25 scientists and engineers and raised a $55M seed round led by Altimeter Capital, Menlo Ventures, Striker VC, and NVIDIA.

The company works on core problems in AI and turns multimodal research into scalable systems for real-world use.

This engineering-first, high-impact role connects core research with production deployments. You will work where product engineering, model evaluation and customer architecture meet, alongside founders, early enterprise partners and the internal research team.

You will turn complex multimodal requirements into robust, high-performance systems and send real-world constraints back into the model roadmaps.

Responsibilities

• Design and ship low-latency, scalable application architectures that bring the model into complex enterprise data pipelines and high-throughput workflows.
• Build open-source reference implementations, production demo apps, developer tooling and SDK components that set the pattern for building on the model.
• Create rigorous evaluation pipelines and benchmarks that compare the model with other frontier multimodal models using varied real-world visual datasets.
• Find and fix bottlenecks across application pipelines, data ingestion, inference APIs, and prompt and orchestration logic.
• Act as lead technical architect for engineering teams at design-partner companies, guiding discovery, proof-of-concept work and rollout to production.
• Turn failure modes, edge cases and performance needs seen in real applications into prioritized specs for research and model infrastructure teams.

Requirements

• 5+ years building and shipping high-performance production software in Python, with clean, maintainable code and deep reasoning about your architectural decisions.
• Ideally, you also work in TypeScript or Go/Rust.
• Experience building complex LLM or multimodal systems beyond simple API calls, such as agentic workflows, structured tool-use frameworks, evaluation pipelines or orchestration layers for vision-language models.
• Hands-on work evaluating, fine-tuning or benchmarking ML models, and telling apart model failures, pipeline errors and data-quality issues.
• A builder's drive to ship quickly, with full lifecycle ownership. Experience as a founding engineer, startup core contributor or forward-deployed engineer is highly valued.
• Familiarity with modern cloud architecture, API design, vector and multimodal indexing, streaming protocols, and high-volume image or video pipelines under tight latency and cost limits.
• Ability to explain architecture tradeoffs clearly to research teams and external engineering leaders.
• Prior experience at a foundation model provider, AI developer platform or ML infrastructure company.
• Direct experience building multimodal applications in vision-heavy fields such as robotics, industrial automation, manufacturing, geospatial, medical imaging or media production.
• Experience building public developer APIs, open-source AI tooling or developer ecosystems.

Conditions

Full-time employment.

Salary of $250,000 to $400,000.

Benefits

• Significant equity.
• Comprehensive health insurance.
• Perks.