Forward Deployed AI Engineer, Travel Intelligence

Hopper
•$360,000 - $425,000•Remote

About The Position

Hopper’s Research team builds travel-specialized models, and the Forward Deployed AI Engineer brings that work into partner environments. The engineer works with partner teams to build evaluations, prepare partner data, train and evaluate models, and deploy them into production to deliver measurable results. As a member of a small forward-deployed team, this role leads model training on site across areas such as servicing, loyalty, and commerce, and serves as a bridge between partners and Research. The role involves developing held-out benchmarks and success measures with partner data and business teams; preparing training data within approved partner environments; conducting post-training with Research; and running jobs on partner or otherwise approved cloud and GPU infrastructure. The engineer designs experiments, moves evidence-backed models into production, monitors their quality, latency, and cost, and uses production results to improve them. They also present and defend results and trade-offs to partner leaders, share field learnings with Research, and create reusable deployment tooling. The position requires careful adherence to data-handling, privacy, and security standards in partner environments, as well as clear documentation to support partner model-risk, security, and compliance reviews. The role requires travel and multi-week on-site partner engagements, and a home base in the Eastern or Central time zone.

Requirements

  • Have independently run the full end-to-end post-training loop, from data preparation through fine-tuning an open-weight model, evaluation, and serving; experience limited to using a vendor’s fine-tuning service is not sufficient.
  • Have working experience with supervised fine-tuning and at least one of preference optimization or reinforcement learning, including parameter-efficient methods such as LoRA.
  • Have strong Python skills and command of standard training tools such as PyTorch, Hugging Face, and vLLM, and be comfortable running jobs on rented or customer GPUs.
  • Have built benchmarks and graders from messy real-world data and used them to compare models on quality, cost, and latency.
  • Use held-out sets, baselines, and ablations, and be able to explain the limitations of results as well as headline findings.
  • Have completed at least two deployments embedded with an enterprise customer, working within their systems, data controls, and security reviews.
  • Be confident presenting to and challenging senior customer stakeholders, including data and technology leaders.
  • Have strong data engineering skills across SQL, pipelines, PII handling, and redaction.
  • Understand privacy, security, and model-risk requirements for regulated customers such as banks and airlines.
  • Be willing to travel and spend multi-week periods on site with partners.
  • Be based in the Eastern or Central time zone with daily overlap across partner and Research teams.

Nice To Haves

  • Forward-deployed, applied research, or applied AI experience at an AI-native company that trains its own domain-specific models.
  • Multilingual training and evaluation experience, or experience with travel, airline, or banking systems and data such as PSS, GDS, NDC, ticketing, loyalty, and contact-center data.
  • Experience with agent systems and tool use, or working alongside a research team to turn its methods into customer results.

Responsibilities

  • Work with partner data and business teams to create a frozen, held-out benchmark for priority use cases.
  • Define success criteria for each task, including task completion, policy adherence, cost, and response time; build automated graders and conduct human spot checks.
  • Establish baselines for fair model comparisons.
  • Prepare data within partner environments, including extraction, redaction, deduplication, labeling, and train/test splitting, using only data needed for agreed tasks.
  • Conduct post-training with Research, applying modern fine-tuning methods to partner use cases.
  • Run training on partner cloud and GPU infrastructure, or in another approved environment, while following security and access rules.
  • Design experiments that clearly demonstrate what drives results.
  • Move models from evaluation into production based on evidence, then monitor quality, latency, and cost.
  • Use production results to continue improving models.
  • Present and defend results, methods, and trade-offs to partner data and technology leaders.
  • Work closely with Research and share learnings from partner deployments.
  • Turn deployments into reusable tooling to accelerate future partner work.
  • Uphold data-handling, privacy, and security standards in partner environments.
  • Support partner model-risk, security, and compliance reviews with clear documentation of data, training, and evaluation.

Benefits

  • Competitive salary and pre-IPO equity packages.
  • Unlimited PTO.
  • Carrot Cash travel stipend.
  • On-demand co-working space through FlexDesk.
  • Work-from-home stipend.
  • Generous parental leave, described as above industry standards.
  • 100% employer-paid medical, dental, and vision coverage for employees.
  • Disability and life insurance.
  • Health Reimbursement Account (HRA).
  • DCA/FSA.
  • Access to a 401(k) plan.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service