Forward Deployed Engineer, Applied AI
Confirmed on the employer's own hiring board on Nov 10, 2026. First seen by Alion on Nov 10, 2026. Pika scores B on the Alion truth index.
About the role
At Pika, we're looking for a technically strong, client-facing AI researcher who can turn customer feedback into models and workflows that solve real business needs. You'll work directly with clients to understand what they want to achieve, identify where current model outputs fall short, and fine-tune or post-train models for their use cases. When the solution calls for more than model adaptation, you'll build the supporting workflow.
This is a full-time, hands-on, revenue-driven applied research role. Clear communication, professional client service, reliable delivery, and independent execution matter as much as technical ability. You'll help clients get value quickly, build their confidence in working with Pika, and help turn successful technical work into closed engagements and ongoing business.
What you'll do
Understand the client's needs. Listen carefully, ask focused questions, and translate requests and feedback into a clear problem, practical scope, and agreed success criteria.
Adapt models for real use cases. Fine-tune and post-train models to address client needs. Prepare relevant examples and evaluation criteria, diagnose gaps, and verify that each iteration improves the outcomes the client cares about.
Build the complete solution. Create or adapt workflows around models when needed, connecting APIs, tools, and processing steps into a usable solution that fits the client's deployment needs.
Iterate quickly with clients. Share working results early, gather specific feedback, and make focused improvements in short cycles. Balance quality, complexity, cost, and delivery speed to keep projects moving.
Communicate professionally. Explain technical options and limitations clearly, set realistic expectations, respond thoughtfully to requests, and keep clients informed about progress and next steps. Build a relationship that makes them want to continue working with Pika.
Own execution and coordinate with the team. Drive day-to-day work with independence while keeping teammates aligned on requirements, priorities, progress, and blockers. Ask for input or escalate risks early when needed.
Deliver for deployment. Help move successful experiments into usable deployments, with clear evaluation results, reproducible workflows, and the documentation needed for handoff.
Help close client engagements. Demonstrate the solution's value, address technical concerns, and work with the team to turn successful iterations into client commitments and revenue.
What we're looking for
Hands-on experience fine-tuning or post-training generative AI models, with the ability to diagnose model behavior and evaluate whether an adaptation solves the intended problem.
Strong coding and practical ML skills: you can implement experiments, work with data, debug model pipelines, and build supporting workflows.
Excellent listening and communication skills. You can understand an ambiguous request, clarify what matters, and explain technical work to clients with different levels of expertise.
Professional judgment, reliability, and a service mindset. You handle feedback constructively, communicate limitations honestly, take ownership of commitments, and build trust through consistent follow-through.
Independence and ownership, combined with strong communication and collaboration within the team.
The ability to work in fast feedback cycles, prioritize the highest-impact improvements, and connect technical decisions to client value and commercial outcomes.
Preferred experience
Prior professional experience in a similar client-facing AI or ML role, with end-to-end delivery from requirements and model adaptation through evaluation, integration, and deployment. This experience is strongly preferred.
Delivering reliable AI solutions for external clients or users and supporting them through feedback and iteration.
LoRA fine-tuning, other parameter-efficient adaptation methods, or feedback-driven post-training.
Building model evaluation pipelines, API integrations, or workflows used in a deployed product.
Participating in customer demos, technical pilots, or the delivery of paid projects.
What success looks like
Success is measured by how well you serve clients, how quickly you turn feedback into useful improvements, and how reliably your work meets their deployment needs. Strong performance builds client trust, helps close engagements, encourages continued collaboration, and contributes to revenue.