Forward-Deployed AI Engineer

H.I.G. Capital•San Francisco, NY

3d•Remote

About The Position

H.I.G. Capital is building a small, high-impact AI team to accelerate value creation across its portfolio companies. This role, Forward-Deployed AI Engineer, will be forward deployed to drive measurable business impact across H.I.G. Capital’s portfolio companies through AI-powered automation. This role sits within Portfolio Operations, reports to the Operating Partner, AI, and works alongside H.I.G.’s Strategy and Transformation Operating Partners, who identify high-value opportunities and direct focus areas. As a Forward-Deployed AI Engineer, you will work directly with portfolio companies to identify, evaluate, implement, and measure AI solutions that generate tangible EBITDA impact. The majority of your work will involve selecting and deploying third-party AI platforms, integrating them with existing systems, measuring their impact, and training teams to use them effectively. You will also build custom solutions when off-the-shelf tools are insufficient. This ranges from lightweight automations built with no-code and low-code platforms (e.g. Claude Cowork, n8n, or Zapier), which are often the fastest and most practical solutions, to fully custom agentic workflows built in LangGraph or equivalent frameworks when the use case demands it. Knowing which approach fits the problem is as important as being able to execute either one.

Requirements

Bachelor’s degree required in Computer Science, Mathematics, Statistics, Data Science, or Engineering.
3+ years implementing, integrating, or building AI or ML systems, with clear evidence of solutions taken from scoping through production deployment and sustained adoption.
Demonstrated experience evaluating and deploying third-party AI platforms in enterprise environments: configuring tools, integrating with existing systems, validating outputs, and measuring ROI.
Technical depth is required not because you will be building every day, but because it is what makes you effective at the core of this role: you need to evaluate whether an AI product actually works, identify when a deployed system is quietly failing, troubleshoot integrations without waiting for vendor support, and train non-technical teams with genuine credibility.
Ability to select and apply appropriate evaluation metrics for a given AI or ML use case; knows how to distinguish signal from noise in model outputs and identify when a strong metric is masking a broken system.
Ability to design rigorous AI evaluations using golden eval sets, LLM-as-judge, human-in-loop review, and A/B testing with appropriate controls; able to apply these methods to vendor-deployed systems, not just custom builds.
Hands-on proficiency with modern AI platforms (Claude Code, Claude Cowork, Codex, Gemini, Cursor, Windsurf, or equivalent) and automation tools (n8n, Zapier, Make, and similar); actively experiments with emerging platforms rather than reading about them passively.
Solid working knowledge of agentic workflow architecture—orchestrator and sub-agent patterns, state management, tool calling, and failure handling—with hands-on experience in LangChain, LangGraph, or equivalent; able to reason through design choices in a technical conversation, not just configure existing templates.
Proficiency with RAG architectures, MCP server integration, API architectures, and webhook-based workflow automation; able to connect AI tools to enterprise systems, diagnose integration failures, and ensure data integrity across system boundaries.
Strong Python skills: able to read, write, extend, and debug production code and AI-generated code; familiarity with cloud platforms (AWS, Azure, or GCP) and standard deployment patterns including containerization and CI/CD.
Working knowledge of common ML model types (clustering, regression, classification, recommendation systems, gradient boosting, deep learning) and evaluation metrics (precision, recall, F1, AUC, RMSE); able to select an appropriate approach for a business problem, evaluate vendor modeling choices critically, and translate model performance into terms a non-technical executive can act on.
Strong business judgment and financial orientation; understands P&L dynamics and the direct link between operational efficiency and EBITDA; able to engage operators and executives credibly and translate technical outcomes into financial impact narratives.
Comfortable in ambiguous, resource-constrained environments; scopes quick wins within 30 to 60 days, sets expectations proactively, and does not overpromise.
Collaborative, adaptable, and discreet; builds trust quickly with portfolio company leadership and operates as a trusted extension of H.I.G. when embedded on-site.

Nice To Haves

Master’s degree in ML, AI, Computer Science, Data Science, or Applied Mathematics preferred.
Non-STEM undergraduate degrees are not a fit for this role.
Experience with n8n, Zapier, Make, or vertical AI SaaS tools is a plus.
Prior experience across multiple companies or industries simultaneously strongly preferred; private equity, management consulting, or forward-deployed engineering is a significant plus.

Responsibilities

Evaluate and select third-party AI platforms and vertical SaaS tools using a structured build-vs.-buy framework; assess vendor architecture, identify failure modes, and make go/no-go recommendations backed by data rather than demos.
Lead end-to-end implementation: configure workflows, integrate with existing enterprise systems (ERP, CRM, HRIS, data warehouses) via APIs, webhooks, and MCP connections, validate data integrity, and manage go-live.
Run structured proofs of concept with upfront success criteria, instrumented measurement, and edge-case testing before committing to full rollout.
Manage vendor relationships and hold partners accountable to delivery timelines and performance targets.
Define baselines and success metrics before every deployment; build evaluation frameworks to monitor output quality, catch silent degradation, and track user adoption over time.
Translate operational results into EBITDA impact and provide data-backed progress updates to H.I.G. Operating Partners and portfolio company leadership.
Deliver measurable results within compressed private equity timescales, typically targeting quick wins within 30 to 60 days alongside longer-horizon transformation projects.
Design and build targeted AI automation solutions using Python, Claude Code, Codex, Gemini, and other leading platforms (e.g. agentic workflows, document processing pipelines, RAG knowledge systems, and intelligent assistants), scoped to fill gaps that vendor products cannot address.
Build lightweight automations using no-code and low-code platforms (Claude Cowork plugins, n8n, Zapier, Make) when simplicity and speed serve the use case better than a custom-coded solution; exercises judgment about when sophistication is warranted and when it is not.
Develop custom Skills, plugins, and MCP integrations to extend third-party AI platforms within portfolio company environments.
Implement prompt engineering frameworks, output validation, and governance controls to ensure production-grade reliability across both custom and vendor-deployed systems.
Apply build knowledge to strengthen vendor work: evaluate whether a vendor’s underlying architecture is sound, identify silent failure modes, and propose improvements that a buyer without technical depth would miss.
Train portfolio company teams on deployed solutions so they can operate, troubleshoot, and extend implementations independently; drive change management to sustain adoption.
Document every deployment with runbooks, user guides, and playbooks that enable cross-portfolio replication without ongoing support.