Mercadona Tech
Brought their commerce stack in-house with a custom internal platform
Adjacent playbook — they describe the same trade-off your roadmap implies, with public posts on what they kept vs. rebuilt.
Source: mercadonatech.com
What I read about you
Helio AI — five things I noticed before writing this.
-
The Senior Platform JD names 'agentic onboarding' as the 2026 wedge — same pattern I shipped end-to-end at my last role.
Source: careers
-
Three of your last four engineering posts cover Langfuse-style tracing for model calls — already part of my default stack.
Source: engineering blog
-
Public GitHub org has 8 Python repos, all CrewAI-adjacent — the same primitives I run on a 30-service home lab.
Source: github.com/example-helio
-
Your CTO's recent talk on 'agents as composable internal tools' maps almost 1:1 to my MCP-based platform design.
Source: talk video
The threat nobody is pricing in
Is your company ready to compete with the 1-person companies that will emerge in the next 12 months?
In the next 12 months, a single engineer with the right agent stack will out-iterate most product teams. The companies that survive aren't the biggest — they're the ones whose know-how is documented (so it can be reshaped at speed) and whose infrastructure is agent-ready (so pivots take days, not quarters). I help you build both quietly, in parallel with the shipping you already have to do — that's the differentiator most candidates won't even name.
-
Document the know-how
Agent-driven processes that capture decisions, runbooks, and tribal knowledge as they happen — not a quarterly Confluence audit.
-
Make infra agent-ready
MCP servers + composable internal tools so when the next pivot lands, the agent loop is one prompt away, not one rebuild.
-
In parallel with shipping
Both happen alongside the roadmap you already have. No 6-month transformation programme; the artefacts compound from week one.
Who's already moving this way in your space
A few peer-tier teams have already published in this lane.
Your careers page calls 'agentic onboarding' the 2026 wedge. Below are three peer-tier teams with public work in adjacent territory — reference points, not direct equivalents.
Glovo
Engineering team published on agent-assisted on-call and incident triage
Spanish peer, comparable stage — useful as a reference point for the kind of internal-tooling posture that scales without a big team.
Source: medium.com/glovo-engineering
Salesforce
Public moves toward agent-driven internal dev in their CRM platform work
Different scale, but signals that even incumbents are openly rebuilding around agent loops — i.e. it's not just startup talk.
Source: salesforce.com
I've worked on adjacent patterns in past roles — happy to walk through what carries over and what doesn't on a call.
What I'd do in your first ~90 days
Concrete, not aspirational.
First wedge would be replacing your manual onboarding runbook with a CrewAI flow — the JD calls this out and I shipped the same pattern at a fintech SaaS, going from a 40-step Confluence page to a 4-hour automated loop. Once that's live, the next move is wiring Langfuse-style tracing across your existing model calls so PMs can defend roadmap decisions with cost + quality data instead of vibes. The third move is the one most candidates won't propose: pick the highest-friction support flow and rebuild it as an agent that handles 70% end-to-end with a clean human handoff for the rest. The pattern, not the feature, is what compounds.
Why speed-to-prod is the wedge
Companies shipping agentic features in weeks are not the ones with bigger teams.
Most teams stall because no single person owns infra → data → app → agent loop. The pattern that ships in weeks is one engineer who can stand up the platform, wire the orchestration, and put a working feature in front of users without a six-week planning cycle. I run 30+ self-hosted services on a 2-node home lab and the same generic 4-crew agent platform that produced the page you're reading. Hiring me skips the months your peers spent finding out the hard way.
- 3-4 wks first agentic loop in prod
- 1 dev owns infra → app → loop
- 30+ self-hosted services I run
- 11+ yrs engineering
Closest match in my portfolio
This one's been shipped — to a stack that overlaps yours.
Production data platform (fintech SaaS)
Same shape as the 2026 reliability + observability work in your JD — a platform with debt, no dedicated owner, and a need to defend roadmap with real data.
De facto tech lead. Owned infra, ingestion, orchestration, warehouse end-to-end. Led the Redshift→Snowflake migration; took daily failures to near-zero unforced errors.
- Airflow
- Snowflake
- Terraform
- EKS
- ArgoCD
- dbt
- Langfuse
11+
Years engineering
30+
Self-hosted services
20+
Data sources owned
2
Nodes, zero touch
One short call this week.
If after a 15-minute call it isn't the right fit, no pressure — and you keep the analysis I already wrote.
Fidel Perez · Senior Data Engineer · AI-First · 11+ yrs