Jobs at Mirage

Build and evaluate agentic ML systems for multimodal creative tasks, especially video. Design training objectives, datasets, fine-tuning strategies, and evaluation frameworks; run experiments, analyze failure modes across planning, tool use, and execution, and iterate toward reliable agent behavior in production.

Develop and scale large-scale video and multimodal generative models, optimize training and inference for low-latency production, implement acceleration techniques (distillation, quantization, pruning), maintain distributed training systems, improve GPU utilization and throughput, and build tooling for experimentation, evaluation, and debugging to move research models into production.

Own the end-to-end recruiting lifecycle for Engineering and Product: source, engage, and hire technical talent; partner with leadership; improve hiring velocity and candidate experience; and recruit diverse candidates.

Design, build, and own backend systems, services, APIs, and data pipelines. Solve scaling, concurrency, and performance challenges. Integrate and operate large generative AI models in production. Instrument and iterate in production; design core platform infrastructure and third‑party integrations.

As a Software Engineer, you'll develop end-to-end experiences for creative workflows, focus on user interface performance, and integrate AI models for enhancing video production. You're expected to own features and manage projects in a fast-paced environment, utilizing full-stack development skills.

Build and ship end-to-end iOS creative workflows, collaborating with designers to create advanced video/audio processing and graphics interfaces, instrumenting and iterating with production data in a fast-paced startup.

Act as embedded product counsel for product, engineering, and research: review features, launches, and model deployments; advise on rights of publicity, copyright, training data, privacy, content moderation, and app store compliance; own consumer-facing legal surfaces and in-product disclosures; triage company legal matters and manage outside counsel; build lightweight legal processes to enable fast product delivery.

Design and build production agentic systems for short-form video: architect context gathering, planning and tool selection; integrate state-of-the-art models; measure and iterate on agent quality using evaluation frameworks and real-world feedback; own end-to-end deployment balancing performance, reliability, and scalability.

The Lifecycle Marketing Lead will manage the customer journey and lifecycle programs for Captions' products, enhancing onboarding and retention through various channels. This includes partnering with multiple teams to optimize campaign performance based on analytics and insights.