Principal AI Architect

Straive; New Jersey; On-site; USD 180,000 - 240,000; Full time

Job summary

Straive seeks a senior AI/ML Architect to design and ship production-grade AI systems at enterprise scale, working with real data and users. You will own open-weight fine-tuning strategies, model routing, and efficient serving architectures across client engagements.

You will lead a large team of engineers and data scientists, building reusable accelerators and reference architectures that speed future deployments. A strong blend of coding and architectural leadership is required.

Qualifications

• 12+ years across software engineering, architecture, data platforms, and AI/ML.
• Experience in a startup or product company with engineering velocity and enterprise process.
• BS/MS/PhD in Computer Science, Engineering, or related technical field.

Responsibilities

• Set reference architectures for deployed AI systems (model routing, retrieval design, tool exposure, orchestration, guardrails, observability, and evals).
• Drive prototypes to production via code reviews, automated testing, quality gates, and release readiness.
• Own fine-tuning strategy (base selection, dataset curation, full vs. LoRA/QLoRA) and serving optimization to beat frontier baselines caseload.
• Lead 30+ engineers/data scientists across client engagements; build accelerators and reference architectures.
• Partner with sales/consulting on workshops, proposals, cost/effort estimation, and technical due diligence.

Skills

Python

Distributed systems

APIs

Containers/Kubernetes

GPU serving

CI/CD

Databricks

Snowflake

Open-Weight fine-tuning

Architectural leadership

Education

BS/MS/PhD in CS/Engineering

Tools

Databricks

Snowflake

Kubernetes

Job description

About Straive

Straive is a global leader in data analytics, AI, and knowledge services, partnering with Fortune 500 enterprises across financial services, insurance, life sciences, and information providers. We build and operate data platforms, AI-enabled solutions, and analytics products that power mission-critical business decisions for our clients worldwide.

About the Role

Most enterprise AI never leaves the pilot phase. Roughly 80% of initiatives fail to deliver promised value, and only ~14% of organizations have scaled agentic systems to real operational use. The root cause is rarely the model—it is uncurated pilot data, underestimated integration layers, and missing evaluation infrastructure.

We close that gap inside our clients' environments at a defensible cost per unit of output. Because economics only work when open-weight models are fine-tuned and served where they outperform hosted calls, this is the most consequential technical hire on our team.

What You Will Own

• Architecture Authority: Set reference architectures for deployed AI systems (model routing, retrieval design, tool exposure, orchestration, guardrails, observability, and evals). Serve as the final escalation point on performance, architecture, and deployment.
• Production Standards: Drive prototypes to production via strict code reviews, automated testing, quality gates, release readiness, and deep integration with clients' systems of record/identity.
• Open-Weight Engineering & Unit Economics: Own fine-tuning strategy (base selection, dataset curation, full vs. LoRA/QLoRA) and serving optimization (routing, caching, quantization, batching, distillation) to ensure tuned models beat frontier baselines cost-effectively.
• Team Leadership & Reusable IP: Lead 30+ engineers/data scientists across client engagements. Build accelerators and reference architectures that make subsequent implementations significantly faster and cheaper.
• Commercial Partnership: Partner with sales and consulting on workshops, proposals, cost/effort estimation, and technical due diligence.

What We Are Screening For

• Production at Scale: Proven track record of designing and shipping production AI systems against real enterprise data and real users (pilots/demos do not count).
• Open-Weight Fine-Tuning: Hands-on experience curating training data, choosing fine-tuning strategies (full vs. LoRA/QLoRA), evaluating against frontier baselines, and serving under real load.
• Cost & Latency Optimization: Measurable reductions in running costs or latency via quantization, distillation, routing, or serving architectures.
• Evaluation Infrastructure: Evals treated as a first-class engineering concern (offline suites, CI regression gates, online monitoring, human review).
• Data Readiness & Governance: Command of enterprise data estates, NIST AI RMF, ISO/IEC 42001, EU AI Act risk tiers, and financial services Model Risk Management (MRM) for GenAI/agentic systems.
• Hands-On Depth: Active coding and architecture capabilities across Python, distributed systems, APIs, containers/Kubernetes, CI/CD, GPU serving, major cloud providers, and platforms like Databricks or Snowflake.

Baseline Qualifications

• 12+ years across software engineering, architecture, data platforms, and AI/ML.
• Experience in a startup or product company (we require engineering velocity alongside enterprise process).
• BS/MS/PhD in Computer Science, Engineering, or a related technical field.
• Note: We do not require or prioritize certifications; we care about what you have shipped.

What This Role Is NOT

• Not a Frontier API Integrator: If your experience is primarily prompt engineering and hosted API calls, this role is not a fit.
• Not a Pure Data or Cloud Architect: Warehouse/pipeline design and cloud infrastructure depth are prerequisites, not the full job.
• Not Classical ML Only: Predictive modeling experience without generative and agentic production experience is insufficient.
• Not a Hands-Off Manager: You must still write and review code.

Equal Opportunity Employer

Straive is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, or veteran status.