Systems

Systems design

Automation that keeps working when the process gets complicated.

I design the full journey: input, orchestration, execution, validation, handoff, and observability. The specific technology can change; the system's reliability should not.

01Input
Documents, APIs, or desktop.
02Orchestration
Priorities, rules, and coordination.
03Execution
Extraction, automation, and decisions.
04Control
Validation and human review.
05Handoff
Context delivered to the right team.
06Observability
Logs, alerts, evidence, and retries.

Monitored flow 6 stages Automatic retry

Capabilities

End-to-end architecture and execution

Build: Python, APIs, OCR, and process automation.
Operate: Workflow orchestration and containerized systems.
Reliability: Retries, idempotency, evidence, and fallbacks.
Collaboration: Translation across operations, sales, onboarding, and technology.

Cases

Anonymized evidence

Three representative proofs from a much wider field of action.

This is a representative sample, not a closed catalogue: only three examples that demonstrate my expertise across very different operations. Sensitive details are intentionally omitted.

CASE 01Operational automation

High-volume execution without turning every exception into manual work.

Problem: Repetitive processes spread across multiple interfaces and tools.
My role: I designed the architecture, code, controls, and production handover, coordinating with the teams involved.
System: Controlled runs with validation, retries, and reviewable evidence.
Outcome: Operational capacity that can scale without manual work growing at the same rate.

Scale signalThousands of repeatable runs · logged failures · retries and reviewable evidence.

CASE 02Document intelligence

Messy documents turned into data ready for operational use.

Problem: Critical information trapped in variable, unstructured formats.
My role: I designed the pipeline, quality controls, and connection to the operational workflow.
System: OCR, extraction, rules, and human review at points of uncertainty.
Outcome: Less manual transcription and clear lineage from source to operational data.

Control signalHuman review is focused on uncertain cases within an aggregate volume of more than 100,000 documents and processes managed.

CASE 03Conversational operations

Conversations that prioritize intent before consuming operational time.

Problem: Too much effort spent on conversations with no fit or next step.
My role: I designed the agent, qualification rules, team handover, and end-to-end experience.
System: Controlled context, filters, scheduling, and escalation only when valuable.
Outcome: Better team focus and more efficient use of automation resources.

Outcome signalThe team receives qualified conversations with context and a defined next step.

Professional briefing

Hands-on technical leadership · Málaga, Spain

Raúl Paredes

Lead AI & Automation Engineer

I design and build AI and automation systems that turn manual, fragmented operations into reliable, traceable workflows engineered for high volume.

Connect on LinkedIn View experience & education

Have a difficult operation to automate? See how I work Explore the CarDreamer case

Production impact

More operating capacity without multiplying manual work: a direct foundation for profitable growth.

OPERATING SCALE 100,000+ executions Processed with control and traceability View evidence CAPACITY RECOVERED 1,000+ hours Of manual work avoided View experience AUTONOMOUS OPERATION <5% intervention Only the cases that genuinely need it View system

Contact

LinkedIn is the best place to start.

For a technical role or a complex operational problem, send the context and I will respond with a clear first read. Málaga, Spain · remote-first approach.

Connect on LinkedIn Send email

Experience

Technical delivery with business context.

Hands-on architecture, code, operations, and cross-functional alignment.

APR 2025
PRESENT

bcount

Lead AI & Automation Engineer

Growth-stage fintech/SaaS backed by a €1M+ seed round in 2026, with a proprietary platform combining advanced automation and AI agents across tax, accounting, and payroll workflows.

Technical owner of business-critical AI and automation systems supporting a core part of the value proposition and allowing the operation to scale without increasing manual work at the same rate.
Identify and prioritise use cases with Sales, Onboarding, Product, and Operations, from business need through production delivery.
Define production controls covering validation, traceability, retries, observability, and failure recovery.
Provide technical mentoring on AI agents, prompt engineering, and safe AI practices.

Scope: architecture · code · orchestration · operations · handoff

OCT 2024
APR 2025

Management Solutions

Data Science & IT Business Consultant

Technical analysis and consulting across data, efficiency, and innovation initiatives.
Translation of business needs into technical decisions and deliverables.
Delivery in demanding, multidisciplinary, client-facing environments.

SEP 2023
AUG 2024

Wains

Data Scientist

Built applied data, machine-learning, and automation systems.
Worked with multimodal information and automated web processes.
Progressed from internship work into full-time technical responsibility.

Education

2022—2024Master's Degree in Computer EngineeringUniversity of Jaén · EQF 7

2018—2022Degree in Computer EngineeringUniversity of Jaén · Information Systems

JAN 2023Entrepreneurship Programme — 1,2,3 Emprende1,2,3 Emprende · January 2023

APR 2022Digital Experience: Data ScienceIMMUNE Technology Institute

Languages

Snake · Lab

Easter egg

Arcade online

Snake

A small classic inside Raul OS. Use arrow keys or WASD.

Score0

Best0

StatusReady

Exit game

Minesweeper · Lab

Raul OS Lab

Minesweeper

Safe first move

Mines10

Flags0

Time0:00

StatusReady

Leave game

Click to reveal · right-click to place a flag.

Tic-tac-toe · Lab

Human vs Raul OS

Tic-tac-toe

Local AI · no network

You0

Raul OS0

Draws0

Your turn · you are X

Leave game