Samuel Beuran

Lead AI Engineer

Lead AI Engineer with 8+ years of experience designing, building, and scaling production-grade data and AI platforms across fintech, e-commerce, and telecom. Deep expertise in Azure-centric data and AI stacks (Databricks, Delta Lake, Azure Data Factory, Airflow, Spark/Scala, Python, SQL), combined with hands-on delivery of GenAI and agentic AI systems (RAG, LightRAG, LLMs) in regulated enterprise environments.

Proven track record delivering end-to-end AI platforms at scale, processing 5TB+ of data and supporting 2M+ transactions per day, preventing €1.2M+ in annual fraud losses and enabling customers to launch AI-driven analytics and financial services generating €1.4M+ in revenue. Designed and operated an Agentic AI Data Platform that ingests large-scale enterprise data via ETL pipelines, transforms it into analytics-ready datasets, and delivers dashboards, insights, and AI-powered recommendations embedded directly into business workflows.

Personal SaaS
PureText.io•1k MRR
See CVDownload CVGet in Touch

Experience

Lead AI Engineer
Nationale-Nederlanden (Fintech)

Agentic AI Fraud Detection / KYC Platform / AI Data Insights Platform

07/2022 - Present

Current
  • • Led the design, implementation, and scaling of production-grade AI platforms used by banking and insurance customers in regulated environments
  • • Designed and delivered an Agentic AI Fraud Detection and KYC Platform on Microsoft Azure using Databricks, Delta Lake, Azure Data Factory, Airflow, Python, SQL, Spark, and Scala
  • • Implemented LLM-powered, agentic AI workflows using RAG, embeddings, and vector databases, integrated with TensorFlow and Vertex AI
  • • Delivered real-time transaction analysis and case scoring for 2M+ transactions/day, achieving sub-300ms P95 inference latency and 99.99% availability
  • • Built and scaled an AI Data Insights Platform ingesting and processing 5TB+ of enterprise data, transforming raw inputs into analytics-ready datasets, dashboards, and AI recommendations
  • • Enabled 200+ enterprise clients to generate AI insights across fraud, risk, and compliance workflows, contributing to €1.4M in revenue generation
  • • Orchestrated data, ML, and LLM pipelines with Databricks, including model retraining, evaluation, drift detection, monitoring, and data quality checks
  • • Designed low-latency, scalable architectures aligned with ISO 27001 and NIS2, enforced via CI/CD-integrated SAST/DAST and dependency scanning (Snyk, SonarQube)
  • • Applied FinOps optimizations, reducing Azure cloud and AI platform costs by €10K+ annually

Tech Stack:

PythonJavaSparkScalaPyTorchTensorFlowLangChainLlamaIndexVector DBsAirflowDatabricksAzure OpenAIVertex AIAzure AI FoundryAzure Cognitive SearchAzure Machine LearningTerraformKubernetes (AKS/EKS)GitHub ActionsSnykSonarQubeGrafana
Senior Software Engineer
Carlsberg (E-Commerce)

08/2020 - 07/2022

  • • Contributed to a new international e-commerce platform serving 1M+ users
  • • Led architecture and development of distributed backend services and secure APIs
  • • Designed and deployed cloud-native microservices with full CI/CD automation, ensuring 24/7 availability
  • • Developed 10+ microservices in Python and Java for order processing, stock updates, checkout flows, and Stripe payments
  • • Designed and implemented Databricks ETL pipelines (PySpark, Delta) ingesting 1–2 TB/month from Kafka, APIs, and SQL sources
  • • Re-architected 100+ CI/CD pipelines with GitHub Actions and Terraform, reducing deployment failures by 90%
  • • Containerized services with Docker and Kubernetes (Istio), achieving 99.98% SLA
  • • Implemented observability with Prometheus and Grafana, reducing page load latency by 35%
  • • Enforced security best practices: encryption, IAM least privilege, vulnerability scanning (Snyk, SonarQube)

Tech Stack:

AzureAWSDockerKubernetesKafkaTerraformPythonDatabricksGrafanaPrometheus
Software Engineer
Truphone (Telco – eSIM Services)

07/2017 - 08/2020

  • • Contributed to cloud-native transformation of global eSIM and GSM platforms serving 1M+ users
  • • Led migration from legacy infrastructure to microservices-based platforms on Azure and AWS, achieving 99.95% uptime
  • • Supported onboarding of 50K+ users/year through scalable backend and data architectures
  • • Rewrote core provisioning, billing, and account services in C# (.NET Core), reducing API latency by ~70%
  • • Designed and implemented 10+ REST APIs for SIM activation, billing, and provisioning
  • • Built multi-region infrastructure using Terraform and GitOps (ArgoCD)
  • • Implemented CI/CD pipelines with GitHub Actions across multi-cloud environments
  • • Designed multi-tenant data models using MySQL and Cosmos DB
  • • Integrated monitoring, alerting, and security controls (Prometheus, Grafana, Fortify, OWASP ZAP, SOC2 logging)

Tech Stack:

AzureAWSDockerKubernetesC#.NET CoreTerraformArgoCDCosmos DBMySQL

About Me

Core Tech Stack

PythonSpark/ScalaDatabricksAzure Data FactoryAirflowTensorFlowLangChainRAGAzure OpenAIKubernetesTerraform
Lisbon, Portugal
+351 936 365 541
samuel.beuran98@gmail.com
samuelbeuran-ai.com

Key Projects

Personal SaaS
PureText.io
1K MRR
AI Text Humanization & Detection Bypass Platform

PureText converts your AI-generated content into fully humanized, undetectable writing—ensuring it passes every AI detection tool. Built with advanced NLP models and machine learning algorithms to create natural, human-like text that bypasses all major AI detectors.

Next.jsTypeScriptOpenAI APIMachine LearningNLPVercelStripe
Impact: 1k MRR, successful SaaS business
Visit Site
AI Fraud Detection Platform
Real-time fraud detection using GPT-4 and ML models

Built a highly scalable fraud detection system processing 2M+ daily transactions with real-time anomaly scoring using GPT-4, embeddings, and vector databases.

PythonGolangKafkaGPT-4TensorFlow
Impact: Prevented €1.2M+ in annual fraud losses
Data Warehouse Platform
ETL pipelines and analytics for 5TB+ monthly data

Designed robust ETL pipelines with Spark, Scala, and Airflow for processing massive amounts of customer and transactional data, enabling BI dashboards and compliance reporting.

SparkScalaAirflowTensorFlowKafka
Impact: Reduced processing failures by 55%, cut processing time by 30%
Microservices Migration
Legacy monolith to cloud-native microservices

Refactored legacy monolith into containerized microservices on Kubernetes with Istio for service discovery, mTLS encryption, and A/B testing capabilities.

KubernetesIstioDockerTerraformArgoCD
Impact: 99.99% uptime for 50k+ req/min, 60% faster provisioning
Security & Compliance Platform
Automated security scanning and compliance

Integrated automated security scans into CI/CD pipeline with SonarQube, OWASP ZAP, and custom LLM-based scripts, ensuring PSD2 and GDPR compliance.

SonarQubeOWASP ZAPOAuth2JWTLLM
Impact: 85% reduction in production CVEs, 92% test coverage

Skills & Technologies

Programming Languages
PythonAdvanced
GolangAdvanced
JavaAdvanced
C#Advanced
TypeScriptAdvanced
Cloud & Infrastructure
AWSExpert
AzureExpert
GCPAdvanced
KubernetesExpert
TerraformExpert
AI & ML Technologies
TensorFlowAdvanced
LangChainAdvanced
OpenAI APIAdvanced
Vector DBsAdvanced
AirflowAdvanced

Certifications

Azure AI Engineer Associate (AI-102)Google AI Professional CertificateNVIDIA Generative AI LLMs (NCA-GENL)Microsoft Certified: DevOps Engineer ExpertAWS Certified DevOps EngineerKubernetes Administrator (CKA)HashiCorp Terraform AssociatePrometheus Certified Associate

Education

BSc Computer Science
Iscte - Instituto Universitário de Lisboa

2017

Languages

EnglishC1
PortugueseC2
ItalianB2
SpanishB1

Get In Touch

I'm always interested in new opportunities and exciting projects. Let's discuss how we can work together!

samuelbeuran-ai.com