visitor@​cyriltovena.dev:~$ whoami

Senior Principal Engineer @ Grafana Labs

✉️ cyril.tovena@gmail.com

visitor@​cyriltovena.dev:~$ cat bio.txt

Senior Principal Engineer at Grafana building AI agents at scale, AI observability products, and the distributed open-source systems behind them. Recent work includes Grafana Assistant, Assistant Investigations, and Grafana Sigil, alongside years of work on Loki and Pyroscope.

I build practical AI systems for real production workloads: context-aware copilots, multi-agent workflows, retrieval pipelines, embeddings-backed search, evaluation loops, and telemetry-rich LLM applications running on scalable distributed infrastructure.

My current focus is the full operating model around AI systems: the agent harness, observability, eval loops, tool traces, and control planes needed to make production AI trustworthy for real engineering and operations teams.

I build production AI systems the way strong infrastructure teams build distributed systems: with observability, rigor, and scale as first principles.

visitor@​cyriltovena.dev:~$ cat ai-focus.txt

LLM agentsMulti-agent systemsAI observabilityOnline evaluationsPrompt and agent versioningTracing and tool telemetryRAGEmbeddingsVector retrievalTool callingMCP-style integrationsPrompt engineeringDistributed systems

visitor@​cyriltovena.dev:~$ cat selected-proof.json

AI products in production

Built Grafana Assistant and Assistant Investigations for real operators working across logs, metrics, traces, profiles, dashboards, and incidents.

AI observability harnesses

Now focused on Grafana Sigil: telemetry-first foundations for traces, evals, cost, quality, and agent behavior at scale.

Distributed systems depth

Years of platform and open-source work across Loki, Phlare, and Pyroscope, building reliable backends for high-scale observability workloads.

Operating at global scale

Grafana Assistant is used by 20k users, supports thousands of customers, and runs across 20 clusters worldwide.

visitor@​cyriltovena.dev:~$ cat thesis.md

AI systems should be treated like distributed production systems: observable, evaluated, versioned, cost-aware, and designed to support real operators under real constraints.

visitor@​cyriltovena.dev:~$ cat leadership.txt

  • Set technical direction across AI product, platform, and open-source initiatives
  • Work with teams worldwide across time zones to align product, engineering, and execution
  • Translate ambiguous AI bets into shipped systems with instrumentation, guardrails, and operating discipline
  • Mentor engineers and raise the bar on architecture, execution, and product sense

visitor@​cyriltovena.dev:~$ cat recent-media.txt

From RCA to Autonomous Ops: The Future of AI in Observability

Big Tent S3E7 · Grafana · Feb 2026

How we built Grafana Assistant: AI development for observability

Grafana video conversation · Feb 2026

Regular conference speaker and podcast guest on open source, observability, AI-assisted workflows, and how to run agentic systems with the same rigor as large-scale distributed software.

visitor@​cyriltovena.dev:~$ ls social/

Resume PDFGitHubLinkedInTwitter