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%.