The Position

OPSWAT is building the internal AI & Data platform that will power every non-engineering function at the company — Finance, HR, Legal, Sales Ops, Marketing, CX, IT, Security, and Supply Chain. We are not buying our way out of this problem with off-the-shelf SaaS; we are building it ourselves, with AI-accelerated development, and we need a leader to own the foundation.

As Director of AI Platform & Data, you will own the AI Platform pillar end-to-end and be accountable to executive leadership for its outcomes. You will run the connective tissue that makes company-wide AI enablement real: the MCP servers and gateway, the agent runtime and orchestration layer, the AI-native data platform that feeds them, the analytics layer that surfaces insight through conversation rather than dashboards, and the governance framework that keeps it all safe. You will lead a multi-team organization, make the platform decisions, set the technical and product direction, and represent this pillar to the CEO, COO, and senior leadership. This role is for someone who thinks like a founder, builds like an engineer, and leads like an operator — someone ready to run an org, not assist in running one.

What You Will Be Doing

• Own the strategy, roadmap, and execution for the AI Platform & Data pillar end-to-end: MCP servers, agent runtime and orchestration framework, AI-native data platform, analytics products, and governance.

• Scale and lead a multi-team organization spanning Orchestration, Data Engineering, Analytics, and Governance; recruit, mentor, and develop engineering managers and technical leads — and raise the bar on the leaders you inherit.

• Build the AI-native data platform — a foundation with CDC pipelines, a semantic layer designed for LLM consumption, lineage, freshness SLAs, and a quality framework that makes agent outputs trustworthy by default.

• Reinvent the analytics function. Move the company from static dashboards to agent-driven insights: natural-language analytics, autonomous deep-dives, anomaly detection, and proactive surfacing of business signals to the teams that need them.

• Define and enforce the architectural patterns that let product teams across Finance, HR, Legal, CX, Supply Chain, and GTM Systems ship AI-powered internal workflows on the shared platform — without rebuilding plumbing every time.

• Own OPSWAT’s private LLM strategy and operating model, including model selection, fine-tuning and post-training, secure deployment, inference infrastructure, evaluation, observability, and lifecycle management. Build private models that can power internal agents, AI-assisted SDLC, and sensitive-data workflows where public or externally hosted models are not appropriate.

• Establish the architecture and decision framework for routing workloads between OPSWAT-hosted models and approved external foundation models based on data classification, security requirements, latency, quality, and cost.

• Build and govern the training and evaluation data pipelines required to improve OPSWAT’s private models, including dataset curation, anonymization, access controls, versioning, benchmarking, red-team testing, and protection against data leakage.

• Partner with Security, Engineering, Legal, and Product teams to define private-model use cases across secure coding, internal knowledge retrieval, cybersecurity analysis, operational agents, and regulated or customer-sensitive workflows.

• Eliminate shadow IT by making the platform the obvious choice: design MCP servers, data connectors, and agent primitives that are faster, safer, and easier than DIY alternatives.

• Establish the governance framework — model access, data classification, audit, evaluation, cost controls — that lets the company adopt AI broadly without compromising security or compliance.

• Drive a development culture that runs ahead of the industry curve on agentic coding, AI-assisted SDLC, and internal tooling velocity — the platform team should be the most productive engineering team in the company.

• Represent the pillar to executive leadership, business function heads, and external partners; own the narrative, the roadmap, and the results.

What We Need From You

• 10+ years in software, data, or platform engineering, with 6+ years leading engineering teams and at least 2 years managing managers.

• Demonstrated track record building and shipping platform products at scale — internal developer platforms, AI/ML infrastructure, modern data platforms, or equivalent.

• Deep hands-on fluency with the modern AI stack: LLMs, RAG, agent frameworks (LangGraph, LangChain, or equivalent), MCP, prompt engineering, and evaluation/observability.

• Hands-on understanding of private LLM deployment and optimization, including open-weight models, supervised fine-tuning, preference optimization, quantization, inference serving, GPU infrastructure, model evaluation, and secure model operations.

• Experience designing hybrid AI architectures that intelligently route requests between private models and commercial foundation models while maintaining security, quality, latency, and cost controls.

• Strong data engineering and analytics foundation: lakehouse architectures (Fabric, Databricks, Snowflake), CDC, streaming, dbt-style transformation, semantic layers, and analytics-grade data quality. You have built data platforms that serve both humans and machines.

• Security and governance mindset — you have built systems that needed to pass audit, not just demos.

• Proven ability to partner with non-technical business functions and translate ambiguous requirements into platform capabilities.

• An innovator’s instinct paired with an operator’s discipline: you bring ideas to the table proactively, then ship them with rigor.

• Excellent written and verbal communication; you can hold the room with the CEO and pair-program with an IC in the same week.

It Would Be Nice If You Had

• Background in cybersecurity, critical infrastructure, or another regulated industry.

• Experience fine-tuning or post-training open-weight models for enterprise, cybersecurity, software engineering, or domain-specific use cases.

• Familiarity with private inference platforms and tooling such as vLLM, SGLang, Hugging Face, NVIDIA NIM, Kubernetes-based GPU serving, or equivalent technologies.

• Experience automating revenue operations, finance, HR, or legal workflows through AI.

• Familiarity with SugarCRM, Salesforce, or comparable enterprise system metadata and integration patterns.

• Experience standing up MCP servers, agent infrastructure, or AI gateways in production.

• A point of view on how AI is changing the way internal engineering teams ship software, and how to bring that perspective to a 700+ person company.