Andres Felipe
Resume

andres felipe orjuela arias

Andres Felipe Orjuela Arias

Data EngineerBogotá, Colombia · remote-friendly

Robust data engineering and scalable pipelines from ingestion to analytics.

why this work

I'm a data engineer in Bogotá. I design ETL pipelines and data models from scratch - diagrams, schemas, and all - for banking and pharma clients, and I build RAG systems that answer from company data instead of model memory.

Download resumeGet in touch
Andres Orjuela working from his data engineering workspace
Workspace

Pipeline

one run, stage by stage

The same five-stage shape I design and run for banking and pharma clients. Click a stage to follow one job from bronze to silver to gold.

Stage 01 · SOURCE

Multi-source & Microsoft Server

STREAMING

Concurrent Python extractors & NiFi processors ingest live Microsoft SQL Server databases, REST APIs, and SFTP drops into a raw immutable landing zone.

Stack & Tools:Apache NiFi · REST APIs · SQL Server · Oracle · S3

Experience

  1. Data Engineer

    InetumBogotá, Colombia - Remote

    Mar 2026 - Presentcurrent
    • Migrating legacy databases and building pipelines mostly from scratch based on customer necessities and business requirements, designing distributed ETL with AWS Glue + PySpark and Control-M into an AWS S3 data lake.
    • Writing robust Python code following PEP 8 standards with CI/CD integration, modular GitHub code structure, and Pytest coverage for all class-based functions.
    • Enforcing predefined schemas, table partitioning, and batch loading on writes, while monitoring and debugging ETL pipelines using AWS CloudWatch.
    • Building PySpark data models from scratch to turn custom business logic into structured Silver/Gold/Platinum layers for strategic analytics over AWS Athena + S3.

  2. Data & AI Consultant

    Independent AI & Data ConsultantRemote

    Nov 2025 - Feb 2026
    • Advise clients on ETL pipeline design and automation, applying Python and SQL to streamline reporting workflows.
    • Designed automated data processes that measurably improved operational efficiency for business stakeholders.

  3. Data Engineer - AI & Cloud Solutions

    GJXBogotá, Colombia - Hybrid

    Feb 2025 - Nov 2025
    • Led end-to-end delivery of AI-powered data pipelines on Azure for a life-sciences / pharma client, integrating LLMs and RAG to sharpen response accuracy in production.
    • Designed the ETL and data models from scratch - architecture diagrams, schemas, and transformation layers - turning high-volume structured and unstructured data into decisions.
    • Owned solution architecture across the full lifecycle - requirements, design, deployment, monitoring - and presented it to execs.

  4. Cybersecurity Competitor & Technical Mentor

    SENA - WorldSkills Lyon 2024Bogotá, Colombia

    May 2024 - Dec 2024
    • Ranked #1 Cybersecurity competitor in Colombia, representing the country at WorldSkills 2024 in Lyon, France.
    • Mentored junior competitors in secure coding, network defense, and incident response methodologies.

  5. Big Data Developer - ETL

    Bluetab (IBM)Bogotá, Colombia - Remote

    Apr 2023 - May 2024
    • Built ELT pipelines on AWS Glue + PySpark and on-prem Hadoop/Control-M for banking and pharma clients, processing large volumes of regulated data.
    • Used Apache Hudi tables to manage partitioned, upserting data on the lake without full rewrites.
    • Designed data models and pipeline diagrams from scratch, turning raw sources into analytics-ready schemas for BI teams.
    • Automated REST API extraction and operational workflows with concurrent Python and Bash automation.
    • Enforced data-quality standards with PySpark/Pandas under strict data-governance and risk-appetite protocols.

  6. Big Data Consultant

    2NVBogotá, Colombia - Remote

    Feb 2022 - Apr 2023
    • Built multi-source ingestion with Apache NiFi (SQL Server, Oracle, SFTP, S3) into HDFS in Parquet and Avro for Hive and Impala.
    • Migrated critical legacy SQL processes to PySpark, boosting scalability and query efficiency.
    • Automated daily reports and scheduled delivery with Apache Oozie, Python, and crontab.
    • Optimized Hive/Impala query performance, reducing latency for Power BI analytical reporting.
    • Designed Hive/Impala schemas optimized for executive dashboards.

Architecture

the calls, and what they cost

An engineer's call only counts when it ships. These are the ones I made - what each one cost, and what it came back with.

  1. Impala over Hive for real-time API queries

    API response times reduced from 15-30s to <500ms.
    context

    NiFi-ingested streams needed to serve downstream APIs at 2B+ records/month. Hive’s batch-oriented MapReduce delivered 15-30s latency.

    decision

    Chose Apache Impala’s MPP architecture for sub-second query latency, enabling real-time decisioning instead of batch windows.

    tradeoff

    Higher memory footprint; required tuning for production stability.

    architectural decision

  2. Hudi tables over raw Parquet for lakehouse

    Data quality incident resolution reduced from hours to minutes.
    context

    Bronze/Silver layers needed ACID transactions and audit trails for regulated banking and pharma environments.

    decision

    Implemented Apache Hudi for upserts, time-travel, and data correction without full reprocessing.

    tradeoff

    Added operational complexity; required team upskilling on Hudi APIs.

    architectural decision

  3. 128MB block compaction strategy for HDFS

    Query performance improved 3-5x by eliminating file listing overhead.
    context

    Small file problem causing NameNode memory exhaustion and degraded query performance.

    decision

    Enforced 128MB target file sizes via Spark compaction; avoided high-cardinality partitioning.

    tradeoff

    Increased write latency during compaction windows.

    architectural decision

  4. YAML standardization for pipeline configuration

    Config-related incidents reduced by 40%; enabled non-engineers to review settings.
    context

    Config sprawl across JSON and properties files caused version control conflicts and review friction.

    decision

    Standardized on YAML with Jinja2 templating and schema validation for all pipeline configs.

    tradeoff

    Required migration effort and validation infrastructure.

    architectural decision

Skills

shipped, not just touched

From daily drivers to tools I've spun up in production or tested hands-on.

Languages

languages.prog
Python
daily driver
SQL
daily driver
PySpark
daily driver
Bash
production
Pytest
daily driver
API REST
production

Distributed & Lakehouse

distributed-lakehouse.prog
Apache Spark
daily driver
Delta Lake
production
Apache Hudi
production
Hadoop · Hive · Impala
production
HDFS · Parquet
production

Data & Schema Modeling

data-schema-modeling.prog
Medallion Architecture
daily driver
Data Modeling
daily driver
Schema Design
daily driver
Architecture Diagrams
production
ERDs
production

Orchestration & Ingestion

orchestration-ingestion.prog
Apache Airflow
frequent
Apache NiFi
production
AWS Glue
daily driver
Control-M
daily driver
Apache Oozie
production

Cloud Platforms

cloud-platforms.prog
AWS S3
daily driver
AWS Athena
daily driver
AWS EC2
production
Azure Storage
production
Azure OpenAI
production
Azure Container Instances
production

AI & RAG

ai-rag.prog
RAG
production
LLMs
daily driver
LangChain
production
LangGraph
production
Azure AI Agents
production

Data & DevOps

data-devops.prog
Git
daily driver
Power BI
production
Tableau
production
PostgreSQL
production
Docker
production
Terraform
frequent
Jenkins
frequent

Data Quality & Delivery

data-quality-delivery.prog
Data Quality
daily driver
Pandas
production
Scrum & Agile
daily driver
Data Governance
production
Stakeholder Management
production
Executive Reporting
production

Certifications

credentials & the work that earned them

Vendor certifications, training, and the AI & governance seminar at the Universidad de Chile.

Azure AI Engineer Associate

Microsoft
azure-ai-engineer-associate.certJun 2025 - Jun 2026

Scrum Fundamentals Certified (SFC)

SCRUMstudy
scrum-fundamentals-certified-sfc-.certMay 2024

SQL Intermediate Certification

HackerRank
sql-intermediate-certification.certMar 2024

Complete Practical Apache Airflow Course

Udemy
complete-practical-apache-airflow-course.certCertificate

International Seminar on AI and Governance

University of Chile
international-seminar-on-ai-and-governance.certInternational Mission · 2 weeks

Secure Development

Net4skills
secure-development.certApr 2026 - Apr 2028

Education

degrees & languages

A software engineering bachelor's in progress, built on the SENA programs where I started. I work in Spanish and English.

B.Eng. Software Engineering

in progress
Ucompensar
Bogotá, Colombia2023 - 2026

International Academic Mission · AI & Governance

international seminar
Universidad de Chile
Santiago, ChileInternational Mission · 2 weeks

Associate · Information Systems Analysis & Development

SENA
Bogotá, Colombia2021 - 2023

Vocational Technician · Computer Systems

SENA
Bogotá, Colombia2019 - 2020

Languages

Spanish (native) · English (B2+ · professional working)
for international teamsES · EN

Projects

real architectures, measured outcomes

Work I've shipped in production - the stack each one runs, and the number it moved.

featured
2024 - Presentnews-data-emotion-analyzer-elt

News Data Emotion Analyzer ELT

A lakehouse platform for multi-source news ingestion, built to produce ML-ready datasets for sentiment and emotion analysis.

architecture

  1. Multi-source ingestion into a Bronze landing layer
  2. PySpark + Delta Lake transformations (Bronze → Silver → Gold)
  3. Airflow DAGs orchestrate and monitor every stage
  4. Gold layer exposes ML-ready features for downstream models

outcome

Medallion architecture on AWS raised data quality and analytical performance for emotion-analysis workloads.

PySparkDelta LakeAirflowAWS
View on GitHub
featured
GJX · 2025ai-powered-pipelines-with-rag

AI-Powered Pipelines with RAG

Production LLM pipelines on Azure for a life-sciences / pharma client, grounding model answers in governed company data instead of raw model memory.

architecture

  1. Designed the ETL, data model, and architecture diagrams from scratch
  2. Structured & unstructured ingestion via Python/Pandas
  3. Retrieval-Augmented Generation layer on Azure OpenAI
  4. LangChain/LangGraph orchestration of agent workflows
  5. Monitoring and iteration in production

outcome

Improved response accuracy and relevance for business-facing AI features, bridged for both technical and exec audiences.

Azure OpenAILangChainLangGraphPython
Bluetab (IBM) · 2023–24banking-etl-modernization

Banking ETL Modernization

Migrated legacy SQL data pipelines to PySpark for a regulated banking client, under strict governance and risk-appetite rules - alongside pharma data work in the same role.

architecture

  1. Designed the target data model and medallion layers before migrating
  2. AWS Glue + PySpark ELT jobs replacing SQL scripts
  3. On-prem Hadoop/Spark jobs scheduled via Control-M
  4. Data-quality checks enforced at every pipeline stage
  5. Git/Bitbucket + Jenkins CI/CD for full-lifecycle delivery

outcome

Processing time cut by 50%, with ingestion 70% faster and team operational workload reduced by half.

AWS GluePySparkHadoopControl-M

Field Notes

notes between commits

Pipelines, Linux, and the tutorials I keep coming back to - written when the compiler gives me a minute.

  1. 01I Turned My Kindle Highlights Into Daily Inspiration - And Built a Tool So You Can TooMay 2025tooling
  2. 02Obtén las herramientas de JetBrains GRATIS con tu correo @soy.sena.edu.coApr 2025community
  3. 03Why Every Programmer Should Try Linux: Lessons from 4+ Years of ExperienceJan 2025linux
  4. 04Emotion Analysis with Python and Google Gemini AIMar 2024ai · nlp
  5. 05Tutorial: Nifi Upload Files to SharePoint API with TemplateFeb 2023integration
source · medium Read all notes on Medium

WorldSkills

Lyon 2024 · ranked #1 in Colombia

WorldSkills is the “Olympics of skills” - the world’s largest skills championship, held every two years since 1950. At the 47th edition in Lyon, 1,400+ competitors from ~70 countries fought across ~60 trades. This page earned Colombia’s seat in Cybersecurity - and returned to mentor the next cohort.

the full record lives on the dedicated story page

rank
#1
country
Colombia
event
WorldSkills Lyon 2024
discipline
Cybersecurity
fig · the record on tape
WorldSkills Lyon 2024 - the arena · official
the arena · official
WorldSkills Lyon 2024 - the road to Lyon
the road to Lyon
WorldSkills Lyon 2024 - back home
back home
WorldSkills Lyon 2024 - back home · team
back home · team
WorldSkills Lyon 2024 - back home · portrait
back home · portrait
WorldSkills Lyon 2024 - the celebration
the celebration
WorldSkills Lyon 2024 - the medal
the medal
Press & coverage5 sources · expand to browse
Open the full WorldSkills story

Awards

the run that led to Lyon

Back-to-back national titles, then the international stage. The competition record behind the WorldSkills gallery above.

  1. International competitorSep 2024

    WorldSkills 2024 · Cybersecurity

    Represented Colombia at the international competition in Lyon, France - one of the most prestigious technical skills competitions in the world.

    Lyon, France
  2. 1st place · National FinalJul 2024

    WorldSkills Colombia · Cybersecurity

    Won the national championship and qualified as Colombia's official representative for WorldSkills 2024.

    Colombia
  3. 1st place · National SemifinalNov 2023

    WorldSkills Colombia · Cybersecurity

    Took first place in the national competition, advancing to the final championship.

    Colombia
  4. 1st place · RegionalMay 2023

    WorldSkills Bogotá · Cybersecurity

    Took first place in the Distrito Capital regional competition, advancing to the national championship.

    Bogotá
  5. 1st place · NationalOct 2021

    SENASOFT 2021 · Cybersecurity

    Won the SENASOFT national cybersecurity championship - the start of the run.

    Colombia

Coding activity

live from WakaTime

Live metrics from WakaTime - no claims here, just what I've actually been building and how much time it takes.

Languages & FrameworksTop technologies used in recent development
Languages & Frameworks
Operating systemsTop Operating systems used in recent development
Operating systems

sourcelive Full WakaTime profile

Contact

say hi, or talk about a role

Open to data engineering, cloud migration, and AI/RAG work. Send a message or use a channel below.

start a conversation

Send a message

Tell me what you're building. I reply from itsfelipe.dev@gmail.com within a day or two.

Send message