# Nick Cerutti, senior AI engineer for production agents and LLM infrastructure

Nick Cerutti is a senior AI engineer and AI architect who builds and operates generative AI and agentic systems in production. Teams bring him in to take agent prototypes to production, to bring AI spend under control, to make model serving faster or cheaper, and to own AI architecture when nobody does.

He has worked in AI infrastructure and MLOps since 2021, first at Iguazio (acquired by McKinsey & Company in 2023) and then as an independent consultant to AI platforms, and works remotely with teams across the Americas and Europe. He is open to full-time Senior or Staff AI Engineer, AI Architect and Head of AI roles, and takes consulting and fractional architecture engagements.

## At a glance

- Role: Senior AI Engineer & Architect · independent consultant since 2023
- Focus: Production agents, LLM inference and serving, AI system design, AI cost and governance
- Experience: Iguazio, AI infrastructure & MLOps (2021-2023) · independent consultant to 6+ AI platforms (2023-present) · quantitative systems engineer on a live trading system (2026)
- Results: $200K/yr aggregate cut in cloud and LLM spend · SLO compliance from 71% to 96% on three platforms · a projected $40K/day agent cost explosion contained within 72 hours · multi-tenant substrate for 250-450 concurrent agent instances
- Stack: Python, Go, Rust, TypeScript, CUDA · vLLM, Triton Inference Server, PyTorch · AWS Bedrock and AgentCore, Azure, Kubernetes · MCP, LangGraph, CrewAI, AutoGen
- Education: B.S. in Computer Science, S21 University (2021) · Certified Kubernetes Administrator (CKA), The Linux Foundation (2023) · MLOps Specialization, DeepLearning.AI (2024)
- Works: Remote · teams across the Americas and Europe · English, Spanish and Portuguese

## Problems he is hired for

Public code shows how he builds and the decisions behind it. Client results carry their scope and limits on each case page. Samples are synthetic or reproducible arithmetic, labelled as such.

### An agent prototype works in demos and now has to run in production.

Usually brought by: Founders and CTOs.

Scoped tool permissions, evaluations in CI, per-session spend caps, tenant isolation and SLOs, so the agent can be trusted with real users and real money. He built this layer for 6+ platforms running 250-450 concurrent agent instances on AWS Bedrock, AgentCore and Azure.

Engagements:

- [Agent Production Audit](https://nickcerutti.com/services/agent-production-audit) (2 weeks, from $12K)
- [Substrate Build](https://nickcerutti.com/services/substrate-build) (6-8 weeks, Quoted on the call)

Evidence:

- [Drako: policy engine and runtime controls for agents](https://github.com/drakolabs/drako) (Public code)
- [Forge: tracing, budgets and memory for LangGraph, CrewAI and AutoGen](https://github.com/angelnicolasc/forge) (Public code)
- [Building multi-tenant AI agent platforms](https://nickcerutti.com/work/substrate) (Client case)

### The AI bill is growing faster than usage, and nobody can say which feature or customer drives it.

Usually brought by: Finance and operations.

Cost attribution per agent, model and tool from your own traces; routing each task to the cheapest model that holds quality; caching fixed at the integration layer; spend caps that stop it coming back. Every finding is priced in dollars. Across engagements the aggregate result is a $200K a year cut in cloud and LLM spend.

Engagements:

- [LLM Cost & Routing Sprint](https://nickcerutti.com/services/llm-cost-sprint) (3 weeks, $15K-20K)

Evidence:

- [Inference economics cost model](https://nickcerutti.com/evidence/inference-economics-sample.md) (Reproducible sample)
- [LLM routing and cache savings estimator](https://nickcerutti.com/tools/llm-routing-estimator) (Free tool)
- [LLM cost optimization across client platforms](https://nickcerutti.com/work/cost-program) (Client case)

### Agents are burning money right now.

Usually brought by: Teams in an incident.

Triage and spend attribution before touching the live system, then circuit breakers and per-session budgets, then the controls that keep it from happening twice. One engagement contained a projected $40K/day cost explosion within 72 hours.

Engagements:

- [Production emergency](https://nickcerutti.com/services/production-emergency) (Days, Priced after triage)

Evidence:

- [Containing runaway AI agent costs](https://nickcerutti.com/work/emergency-40k) (Client case)
- [Agent burn calculator](https://nickcerutti.com/tools/agent-burn-calculator) (Free tool)

### Self-hosted models are too slow or too expensive to serve.

Usually brought by: ML platform teams.

vLLM and Triton serving, dynamic batching and quantization, KV-cache layout and eviction, routing between serving tiers, and scheduling for reasoning models. At Iguazio he operated multi-GPU Triton Inference Server clusters and defined their inference SLOs.

Engagements:

- [LLM Cost & Routing Sprint](https://nickcerutti.com/services/llm-cost-sprint) (3 weeks, $15K-20K)

Evidence:

- [Meridian: phase-aware vLLM scheduler for reasoning models](https://github.com/angelnicolasc/meridian) (Experimental code)
- [Stratum: dual-tier LLM serving router](https://github.com/angelnicolasc/stratum) (Public code)
- [KV cache layout after MLA](https://nickcerutti.com/notes/kv-cache-block-layout-after-mla) (Engineering note)
- [Iguazio: inference platform engineering](https://nickcerutti.com/work#iguazio) (Employer engagement)

### The team is shipping AI and nobody owns the architecture.

Usually brought by: CTOs and engineering leaders.

Architecture decisions made, written down and defended; evaluation and SLO frameworks that give the team a shared definition of working; stack and vendor choices with the trade-offs written out; and a hiring specification for the people who take it over.

Engagements:

- [Fractional AI Architect](https://nickcerutti.com/services/fractional-ai-architect) (3 months minimum, $9K-15K/mo)
- [Reliability Foundation — SLOs and evals in CI](https://nickcerutti.com/services/reliability-foundation) (3 weeks, $12K-18K)

Evidence:

- [SLO frameworks for three platforms](https://nickcerutti.com/work#slo-program) (Client case)
- [Phase-aware scheduling for reasoning models](https://nickcerutti.com/notes/phase-aware-scheduling-for-reasoning-models) (Engineering note)

### An auditor, an insurer or the EU AI Act is asking questions the team cannot answer yet.

Usually brought by: Compliance, legal and insurers.

An agent inventory, least-privilege enforced at the tool layer, attributable audit trails and human review where it is needed, packaged as technical evidence a reviewer can check. Technical readiness, not legal advice.

Engagements:

- [Agent Governance & Insurability Package](https://nickcerutti.com/services/agent-insurability) (3-4 weeks, $18K-30K)
- [AI Compliance Package — EU Article 50 plus US module](https://nickcerutti.com/services/ai-act-article-50) (3 weeks, $15K-25K)

Evidence:

- [Agent controls evidence dossier](https://nickcerutti.com/evidence/agent-controls-sample.md) (Synthetic sample)
- [Drako: deterministic controls and audit trails](https://github.com/drakolabs/drako) (Public code)
- [MCP server security checklist](https://nickcerutti.com/tools/mcp-security-checklist) (Free tool)

## Open-source work

- [Graymatter](https://github.com/angelnicolasc/graymatter) (Go): Persistent memory for AI agents with hybrid retrieval and an MCP server, in a single binary. 480+ GitHub stars (October 2026).
- [Drako](https://github.com/drakolabs/drako) (Python): Security and governance for production AI agents: static analysis, policy as code and deterministic runtime enforcement.
- [Meridian](https://github.com/angelnicolasc/meridian) (Rust): Phase-aware inference scheduler for reasoning models on vLLM, with dual-queue dispatch and phase-aware KV-cache eviction.
- [Stratum](https://github.com/angelnicolasc/stratum) (Python): Dual-tier LLM serving with sub-millisecond complexity routing between vLLM and llama.cpp and live VRAM monitoring.
- [Forge](https://github.com/angelnicolasc/forge) (Python): Harness for LangGraph, CrewAI and AutoGen systems that adds tracing, per-agent cost tracking, budgets and memory.
- [Epica](https://github.com/angelnicolasc/epica) (Rust): Rust runtime for formal belief revision in LLM agents, with auditable memory and an MCP server.

## Ways to work together

- Full-time role: Open to Senior or Staff AI Engineer, AI Architect and Head of AI roles, remote. [Download the resume](https://nickcerutti.com/Nick-Cerutti-Resume.pdf)
- Consulting and fractional architecture: Fixed-scope engagements from a few days to eight weeks, and fractional architecture from three months. A $2,500 paid pilot is credited in full. [Services and pricing](https://nickcerutti.com/services)
- Production emergency: Agents burning money right now: email directly and skip the calendar. [n@nickcerutti.com](mailto:n@nickcerutti.com)

## Questions

### Who is Nick Cerutti?

Nick Cerutti is a senior AI engineer and AI architect for production agents and LLM infrastructure. He builds multi-tenant agent platforms, inference serving and the controls around them, and maintains open-source projects including Graymatter, Drako and Meridian.

### What problems is he usually hired for?

Taking an agent prototype to production; an AI bill growing faster than usage; agents burning money during an incident; self-hosted models that are too slow or too expensive to serve; a team shipping AI with nobody owning the architecture; and auditors, insurers or the EU AI Act asking for technical evidence.

### Our AI application costs too much. Can he find out why?

Yes. The LLM Cost & Routing Sprint attributes spend per agent, model and tool from your own traces, prices every finding in dollars, and fixes the largest lines through routing, caching and spend caps. It can start as a $2,500 paid pilot. Across engagements the aggregate result is a $200K a year cut in cloud and LLM spend; results depend on the workload.

### Can he take an AI agent prototype to production?

Yes. That means scoped tool permissions, evaluations in CI, spend caps, tenant isolation and SLOs before real users arrive. The usual starts are a two-week Agent Production Audit or a six-to-eight-week platform build handed over to your team.

### Is Nick Cerutti available to hire as a senior AI engineer?

Yes. He is open to full-time Senior or Staff AI Engineer, AI Architect and Head of AI roles, and takes consulting and fractional architecture engagements. The quickest route is a 15-minute call or an email to n@nickcerutti.com.

### Does he have hands-on LLM inference experience?

Yes. At Iguazio he operated multi-GPU Triton Inference Server clusters with dynamic batching and quantization. In open source he builds Meridian, a phase-aware vLLM scheduler for reasoning models (its real-vLLM benchmark was marked unverified at the September 2026 review), and Stratum, a dual-tier serving router.

### How can I verify his work?

Open the code: Graymatter, Drako, Meridian and the other projects are public on GitHub with their methods and limits. The engineering notes show sources and reasoning. Each case page states its scope and limits.

### How much does an engagement cost?

Prices are published. Work can start with a $2,500 paid pilot that is credited in full if the larger engagement goes ahead. Fixed-scope engagements run from a few days to eight weeks, and fractional architecture starts at three months. The pricing page lists each engagement.

### Does he work remotely?

Yes. He works remotely with distributed engineering teams across the Americas and Europe, in English, Spanish and Portuguese.

## Contact

[Book a 15-min call](https://calendly.com/nickcerutti/15-min-meeting) · [n@nickcerutti.com](mailto:n@nickcerutti.com) · [LinkedIn](https://www.linkedin.com/in/nickcerutti) · [GitHub](https://github.com/angelnicolasc)

Canonical page: https://nickcerutti.com/about