See how candidates think and approach the work this role demands, before the phone screen. We'll build a video challenge from this posting, and you can edit or share it before it goes live.
Pika is hiring applied researchers whose explicit goal is to turn generative models into viral creative experiences that drive traffic and revenue, signaling that consumer AI video companies are shifting from model-building toward growth-oriented, shareable product experiences.
Open to full-time hires and interns.
At Pika, we're looking for a creative, hands-on AI researcher who can turn the latest models, APIs, and emerging ideas into compelling experiences people want to try and share. You'll combine models across image, video, audio, and language to build workflows, demos, and features that bring people to Pika and contribute to revenue growth.
This is a goal-driven applied research role, open to both full-time hires and interns. You'll take ideas from an initial opportunity through experimentation, implementation, and delivery, then use feedback and results to decide what to improve next. Strong creative judgment, resourcefulness, and fast execution are central to the role.
Find promising opportunities. Stay close to new models, tools, creative formats, and cultural trends. Identify ideas with a clear audience, use case, and potential to drive interest, website traffic, or revenue.
Build end to end. Combine existing models, APIs, and tools into working multimodal workflows. Own the path from a rough concept to a usable demo, creative asset, or product experience.
Choose the right amount of research. Test practical approaches quickly. Use lightweight LoRA fine-tuning or targeted post-training when it meaningfully improves the result and fits the project's time budget.
Deliver frequently. Work in short iteration cycles, set a clear scope for each experiment, and get strong first versions out quickly. Make thoughtful trade-offs between additional research, output quality, cost, and time to delivery.
Create work ready to share. Produce compelling demos and creative outputs that the team can use for social media, launches, and other growth efforts. Personally posting on or managing social channels is not required.
Learn from results. Use audience response, product feedback, and relevant traffic or revenue signals to decide which ideas to refine, expand, or stop.
Make successful work reusable. Leave working workflows, examples, and concise documentation so the team can reproduce and build on what you deliver.
Hands-on experience building with generative AI models, APIs, or creative workflows, with examples of work you have taken from idea to a working result.
Strong prototyping and coding skills: you can connect tools, debug a pipeline, and make an end-to-end experience work.
A creative, entrepreneurial mindset and good judgment about what makes an output compelling, useful, and worth sharing. You enjoy the pace and resourcefulness of a hackathon: testing ideas quickly and getting a working result into people's hands.
Curiosity across modalities and the ability to learn unfamiliar models and tools quickly.
Ownership and comfort with tight timelines: you can prioritize, simplify an approach, and recognize when shipping and learning is more valuable than further polishing.
A clear connection between technical choices and the intended user or business outcome.
Experience as a startup founder or early-stage startup team member, or actively building and testing startup ideas.
Hackathon wins or standout hackathon projects that demonstrate rapid prototyping, resourcefulness, and end-to-end execution.
LoRA fine-tuning, lightweight model adaptation, or other targeted post-training.
Combining image, video, audio, or language models in a single workflow.
Building creative demos, interactive prototypes, or AI experiences that reached real users.
You deliver a steady stream of compelling, usable experiments; turn the strongest ideas into repeatable workflows; and use each release to learn what attracts users and contributes to Pika's growth. Success depends on the quality, speed, and impact of what you deliver.
pika:b075e6fb-df01-4eac-9437-1c1db97ea881pikaThis posting hasn't been probed by our closure verifier yet. Stream C runs on a rolling schedule against postings approaching the close-decision threshold.
See how we measure for definitions, or our corrections log for known issues. Found something wrong? Flag a correction.