Applied AI Researcher (Optimization & Domain Models)

This is a remote position.

Role Type: Research-to-Production Specialist Focus: Domain-Specific Models, Optimization & Safety Reports to: Chief AI Evangelist & Product Head

Role Summary

The Applied AI Researcher (Optimization & Domain Models) is responsible for adapting and optimizing AI models for micro-industry-specific problems. This role bridges theoretical rigor and real-world deployment, ensuring models are not only accurate but safe, explainable, and production-ready.

You will own problem-specific reasoning models, vertical LLM tuning, anomaly detection, and quality models that power defensible SaaS offerings.

Key Responsibilities

Domain Model Development

· Design problem-specific reasoning and optimization models.

· Fine-tune vertical LLMs for industry-specific language, workflows, and constraints.

· Build anomaly detection, prediction, and quality inspection models.

· Adapt foundation models to operate under domain rules, policies, and regulations.

Evaluation, Guardrails & Safety

· Own evaluation loops (offline, online, human-in-the-loop).

· Design guardrails for hallucination control, bias mitigation, and policy compliance.

· Implement safety tooling for enterprise-grade AI deployments.

· Define success metrics tied to business and operational outcomes.

Research to Production

· Convert research prototypes into deployable, scalable micro-industry models.

· Partner with engineers to integrate models into agents and SaaS workflows.

· Document model behavior, assumptions, and failure modes.

· Create repeatable model adaptation playbooks.

IP & Thought Leadership

· Contribute to proprietary model architectures and training strategies.

· Publish internal whitepapers and external POVs where appropriate.

· Support GTM narratives with credible technical depth.

Required Qualifications

· Desirable PhD in AI, ML, Applied Mathematics, Operations Research, or related field.

· Strong background in optimization, probabilistic modeling, or deep learning.

· Experience fine-tuning LLMs or training domain-specific models.

· Hands-on experience with Python, PyTorch, TensorFlow, or JAX.

Preferred Qualifications

· Experience with enterprise or regulated domains (healthcare, finance, telecom).

· Familiarity with reinforcement learning or constrained optimization.

· Exposure to safety, alignment, or AI governance frameworks.

Success Metrics

· Deliver 1 domain-tuned model per quarter.

· Demonstrate measurable performance lift vs baseline models.

· Deploy models into at least 2 production workflows.

· Reduce inference errors or quality issues by ≥25%.