LogoAutomation and full-stack developer based in Pampanga, near Metro Manila, Philippines

I’m Allan. I automate workflows and solve real business problems.

I build data pipelines, business applications, and AI automations that turn manual work into reliable systems.

At a glance

Selected work

  • Client project

    Recall campaign site + VIN scraper

    Parse VIN-specific recall data from automotive sites.

    Result Scraped over 1,000,000 records of VIN-specific recall data.

    See how it works: Recall campaign site + VIN scraper
  • Demo system

    Invoice / AP validation

    Invoices arrive as PDF text, not structured input, and a duplicate must never be paid.

    Result In the synthetic test batch, 5 of 6 invoices are routed to a human with a clear reason, and the 1 clean invoice is auto-approved.

    See how it works: Invoice / AP validation
  • Client project

    Legal Scraper & OpenAI Analysis

    Extract case PDFs from legal sites for OpenAI-based analysis.

    Result Saves about 2 days of manual work every week.

    See how it works: Legal Scraper & OpenAI Analysis
  • Own build

    Content studio + distribution

    Hosted render APIs blew the budget, and nothing may go public without an approval gate.

    Result Estimated running cost: Claude (pennies) + ElevenLabs (~$11) + fal.ai (a few $) + self-hosted FFmpeg ($0) ≈ $11–15 a month. In the 6-video test batch, 3 are postable (10 platform posts).

    See how it works: Content studio + distribution

What clients say

  • “Recommended to everyone, the instructions, explanations and directions are very clear. 5 stars! for the service. I'll rehire this person again for our future projects.”

    Upwork client · Sep 28–29, 2026
  • “Allan has been a pleasure to work with! The quality of his work as been great and his communication is clear.”

    Upwork client · Jun 29–Jul 2, 2026
  • “I've used Allan several times now and he's always been professional, on-time and gives me excellent service. I highly recommend him.”

    Upwork client · Jul 8–11, 2026

All client feedbackBook a free consultation

“Allan was very professional, ahead of schedule and I only had to tell him once what I needed. …”

Excerpt · Upwork client, “MLS Script - Modification and Excel Cleanup”, May 23–26, 2026 · All client feedback
  1. 01 Introduction

    Career 1 of 7: Data Scraper / Full Stack Developer, Freelance, via Upwork

    Remote clients, from my base in Metro Manila

    02 My role

    Builds and maintains Python scraping pipelines for legal, financial and automotive datasets, plus full-stack web applications, for independent clients.

    04 Overview · automotive data

    Recall Campaign Site

    1. Celery and Redis run the scraping as scheduled background jobs
    2. the scraper collects VIN-specific recall data from automotive sites, solving captchas during the run
    3. results are exported as CSV files
    4. a Flask admin dashboard monitors the scripts and keeps the historical records
    Explained, not a screen · parts from the project notes · their order is reconstructed
    See how it works in Selected Work

    05 Contributions

    • Full-stack development, API integrations and deploying scraping pipelines.
    • Debugged and enhanced a Flask-based web application.
    • Automating post-analysis workflows with OpenAI.

    Tools: PythonSeleniumGoogle Cloud StorageOpenAI APIFlaskCeleryRedisReactDjangoPostgreSQL

    06 Projects · 8 · Freelance, via Upwork

    • Zdigitech Engineers LMS

      Client project

      One shared space for students, children, parents and instructors.

      Built A multi-role learning management system: lessons, progress tracking and communication between all participants.

      Tools: React · Django · PostgreSQL · Python · LMS · Full Stack Development

    • CI/CD Integration Pipeline

      Client project

      Deployment automation for a client.

      Built A minimal CI/CD workflow with Docker and GitHub Actions.

      Tools: Python · Docker · GitHub Actions · CI/CD

    • Legal Scraper

      Client project

      Case PDFs published on legal websites, collected into cloud storage.

      Built A Python Selenium script that extracts the PDFs and uploads them to Google Cloud Storage automatically.

      Tools: Python · Selenium · Requests · Google Cloud Storage · Automation

    • Till Payments Integration in Flask Application

      Client project

      Payments inside an existing Flask application.

      Built A payment module for PayPal and credit-card payments.

      Tools: Python · Flask · PayPal API · REST · Frontend Integration

    • Shopee Scraper

      Proof of concept

      Automotive product listings from Shopee.

      Built A scraper that captures costs and descriptions into a MySQL database.

      Tools: Python · Selenium · Shopee · MySQL · Data Extraction

    • BarChart.com Automation

      Client project

      Daily stock market data from BarChart.com.

      Built Selenium and Requests extraction saved as CSV, run daily by Windows Scheduler.

      Tools: Python · Selenium · Requests · CSV · Windows Scheduler

    • Loan Management System

      Client project · proof of concept

      Loan management in one web platform.

      Built A full-stack platform: React frontend, Django backend and PostgreSQL, on Ubuntu Server.

      Tools: React · Django · PostgreSQL · Python · Ubuntu

    • Recall Campaign Site

      Client project

      VIN-specific recall data from automotive sites.

      Built A Flask app with background scraping (Celery, Redis), anti-captcha solving, CSV generation and an admin dashboard for monitoring and history.

      Tools: Python · Flask · Celery · Redis · Selenium · Anti-Captcha · CSV

    Other contracts · titles as listed on Upwork · remote

    • Software Engineering Jobs
    • Software engineer jobs around australia
    • Dev/Tech all-rounder with Claude / AI coding
    • Reddit Scraping & Automation Expert

    Freelance · remote, via Upwork

    United States

    • MLS Scraper Modification and support2026Maintained and modified existing MLS scrapers, providing ongoing support.
    • Pines Federal2024Developed Python scrapers for data extraction.

    Located at country level · no city is marked

    Freelance · remote, via Upwork

    Australia

    • Python Full Stack Developer2025–2026Worked on billing functionality and other website enhancements.

    Located at country level · no city is marked

    Freelance · remote, via Upwork

    Guam

    • Loan Management System2025Developed a full-stack loan management platform using React (frontend), Django (backend), and PostgreSQL on Ubuntu Server.

    Located at territory level · no city is marked

    Freelance · remote, via Upwork

    United Kingdom

    • Reddit Scraper for Market Research: Export to Google Docs

    Located at country level · no city is marked

    Freelance · remote, via Upwork

    Europe

    • CI/CD Integration PipelineHelped a client set up a minimal CI/CD workflow using Docker and GitHub Actions for deployment automation.

    Located at region level · no city is marked

  2. 01 Introduction

    Remote

    Career 2 of 7: Python Developer, Shooju Global

    Houston, Texas, US · company headquarters

    02 My role

    Wrote Python scrapers for a data-analytics platform and owned the pipeline from extraction to clean CSV or database output.

    03 Overview

    From extraction to clean output

    1. a source page with the fields to collect
    2. each field is extracted
    3. the fields become one clean row of a table (CSV or database)
    Illustration · sample field names, not client data · the output format (CSV or a database) is from the role

    04 Contributions

    • Scraped static and dynamic websites, APIs and HTML with Beautiful Soup, Scrapy and Selenium.
    • Handled scraping challenges including proxy rotation, CAPTCHA bypassing and rate limiting.
    • Cleaned, processed and structured the data into CSV files or databases.

    Tools: PythonScrapySeleniumBeautiful SoupRequestsPandasNumPy

  3. 01 Introduction

    Hybrid

    Career 3 of 7: Developer, SM Supermalls

    IKEA building, Diokno Boulevard, Pasay

    02 My role

    Debugged and developed internal enterprise applications on the OutSystems low-code platform.

    03 Contributions

    • Worked with cross-functional teams to deliver new features and improve existing systems.
    • Contributed to the continuous enhancement of enterprise applications through critical analysis and rapid deployment.

    Tools: OutSystems

  4. 01 Introduction

    Hybrid

    Career 4 of 7: Computer Services Programmer A, National Electrification Administration

    NEA Building, 57 NIA Road, Diliman, Quezon City

    02 My role

    Built internal tools and proof-of-concept systems for NEA's digital transformation.

    03 Overview · actual screens

    Electronic Budget Utilization System title screenElectronic Budget Utilization System title screen
    Actual screen · Electronic Budget Utilization System (e-BUS)
    Learning Management System login screenLearning Management System login screen
    Actual screen · Learning Management System login

    04 Contributions

    • Moved manual processes onto internal systems as part of NEA's digital transformation.
    • Supported CMS-based system deployment and general software development tasks.

    Tools: PythonDjangoMS SQL

    05 Projects · 3 · National Electrification Administration

    • Actual screen: Electronic Budget Utilization System (POC)
      Electronic Budget Utilization System (POC)

      Internal proof of concept

      Replace NEA's manual budget tracking and approval.

      Built A Django and MS SQL prototype.

      Tools: Django · MS SQL · Python

    • Actual screen: Learning Management System
      Learning Management System

      Internal tool

      Training schedules, quizzes and reports for electric cooperatives.

      Built A Django system with forms and a quiz module.

      Tools: Django · Forms · Quiz Module

    • Raffle App for Events

      Internal tool

      Automated raffle draws at NEA events.

      Built A Python desktop application with a GUI, backed by MS SQL.

      Tools: Python · MS SQL · Desktop GUI

  5. 01 Introduction

    Hybrid

    Career 5 of 7: Information Systems Analyst I, PAGASA

    Science Garden Compound, Diliman, Quezon City

    02 My role

    Led internal tools for weather data, including the PAGASA Data Monitoring Tool for real-time data visualization.

    03 Overview · actual screen

    ICT Management System: semi-annual maintenance table and the change logICT Management System: semi-annual maintenance table and the change log
    Actual screen · ICT Management System · 2020

    04 Overview · automations

    Explained · no screen

    Daily weather updates

    1. Weather APIexternal data
    2. PythonRequests
    3. Text filesthe update text
    4. FTPfile transfer
    5. 4AM · 4PMscheduled
    6. Facebook pageofficial PAGASA page

    Explained · no screen

    CSV generation and FTP transfers

    1. Pythonscripts
    2. CSV filesgenerated
    3. FTPautomated transfer

    05 Contributions

    • Automated CSV generation and FTP transfers in Python.
    • Supported website maintenance and wrote scripts to streamline operational workflows.
    • Assisted in ArcMap GIS data integration, technical support and documentation.

    Tools: PythonDjangoPostgreSQLArcMap

    06 Projects · 2 · PAGASA

    • Actual screen: ICT Management System
      ICT Management System

      Internal tool

      Track the ICT department's IT assets, issues and ongoing technical problems.

      Built A Django and PostgreSQL system.

      Tools: Django · PostgreSQL · Python

    • Daily Weather Automation

      Internal automation

      Weather updates for the official PAGASA Facebook page.

      Built API data turned into the update text with Python and transferred by FTP.

      Result Scheduled to post at 4AM and 4PM.

      Tools: Requests · FTP · API

  6. 01 Introduction

    Hybrid

    Career 6 of 7: Science Research Specialist I, PAGASA

    Science Garden Compound, Diliman, Quezon City

    02 My role

    Helped develop the wind profile application, a Tkinter and ArcMap desktop tool for tropical cyclone bulletins.

    03 Contributions

    • Automated CSV generation and FTP transfers using Python, and integrated data with ArcMap for GIS visualization.
    • Maintained documentation and supported teammates with technical concerns.

    Tools: PythonArcMap

    04 Projects · 1 · PAGASA

    • Wind Profile Application

      Internal tool

      Automate tropical cyclone bulletin datasets and estimate typhoon damage per household.

      My part A Python desktop app (Tkinter, ArcMap) that interpolates the data and visualizes impact predictions. (I helped develop it)

      Tools: Python · ArcMap · Interpolation · CSV · Automation · FTP

  7. 01 Introduction

    Career 7 of 7: Technical Support Representative, Teletech

    Robinsons Novaliches, Quezon City

    02 My role

    Provided technical support for Verizon customers across internet, cable and phone services.

    Gave step-by-step troubleshooting and device configuration help.

Pampanga, Philippines · where I live

04 case studies · problem → result → how it works

Selected work

From where the work was done to what it does: each case starts with the problem and the result, then shows its process.

  1. 01 · Automation · rules + Claude · Demo system

    Invoice / AP validation

    Demo system · synthetic data

    Problem Invoices arrive as PDF text, not structured input, and a duplicate must never be paid.

    What it does · what I built A 13-node n8n workflow: a manual trigger with mock invoices, or a Gmail trigger, then Extract PDF Text and normalize. Rules pre-flag each invoice; Claude extracts fields and returns auto_approve, needs_review or hold, an owner and a vendor-query draft. Same backbone, second job: a 12-node support-ticket triage.

    Result In the synthetic test batch, 5 of 6 invoices are routed to a human with a clear reason, and the 1 clean invoice is auto-approved.

    How it works · step by step

    Demo system · synthetic invoices · step names from the n8n workflow export

    Input · Gmail

    Gmail trigger

    Live path: watches a labelled inbox for invoice emails with PDF attachments.

    Input · n8n

    Mock invoices

    Demo path: six synthetic invoices with planted edge cases, same schema.

    Step · n8n

    Extract PDF text · normalize

    PDF text becomes the same invoice fields the demo path uses.

    Normalized · synthetic
    vendor   Brightline
    invoice  INV-4120
    po       (missing)
    tax_id   (missing)

    Rule · JavaScript

    Rule pre-flag

    Deterministic checks before any AI: duplicate, PO match, missing fields, approval threshold, currency, math.

    Flags raised · synthetic
    INV-1001 (resubmitted) duplicate of INV-1001
    INV-2050               bills $9,800 vs PO $9,000
    INV-3300               over the $10,000 limit
    INV-4120               missing PO and tax ID
    INV-5090               EUR (expected USD), total does not add up

    AI · Claude API

    Claude: extract and decide

    Reads document-only invoices into fields, labels each one, names an owner and drafts a vendor query when information is missing. An invoice with a hard flag is never auto-approved.

    auto_approve 1

    • INV-1001 · Clean invoice

    needs_review 3

    • INV-2050 · PO amount mismatch
    • INV-3300 · Over threshold
    • INV-5090 · Currency and math error

    hold 2

    • INV-1001 (resubmitted) · Duplicate
    • INV-4120 · Missing PO and tax ID

    Step · n8n

    Build outputs

    One run, three payloads. While DRAFT_MODE is on, vendor queries go to a review inbox, not to vendors.

    Output · Slack

    Slack: exceptions

    The invoices a person must touch, with the reason and the owner.

    Alert · synthetic
    5 of 6 invoices need a person
    INV-2050 → Procurement: bills $9,800 vs PO $9,000

    Output · Google Sheets

    Google Sheets: AP ledger

    Every invoice and its decision: the audit trail.

    Ledger rows · synthetic
    INV-1001 · Acme · $4,200 · auto_approve
    INV-2050 · Globe · $9,800 · needs_review

    Review · Gmail

    Gmail: vendor queries

    Follow-ups for a missing PO or tax ID, sent to a review inbox while DRAFT_MODE is on.

    Draft · synthetic
    [DRAFT] Regarding invoice INV-4120
    Please send the PO number and tax ID.

    Result Routing, not payment: the workflow labels, logs and escalates invoices. It never pays one.

    More on this case: Invoice / AP validation

    Constraints

    • Never pay a duplicate
    • Auditable rules a client can read line by line
    • PDF text, not structured input

    Key decision

    Chose: Rules before the model: objective signals computed by code, Claude only judges and writes

    over An LLM-only classifier

    because Rules detect, AI communicates

    Planted edge cases in the synthetic batch
    CaseInputExpected
    Clean invoiceMatches its POauto_approve
    DuplicateInvoice already seenhold
    PO amount mismatchAmount differs from the PO beyond tolerance (larger of $50 or 2%)needs_review (Procurement)
    Over thresholdAmount above the $10,000 default thresholdneeds_review (AP Manager)
    Missing PO and tax IDRequired fields absenthold, vendor query drafted
    Currency and math errorCurrency mismatch and subtotal + tax does not add upneeds_review

    n8nClaude APIGmailSlackGoogle Sheets

    Discuss a similar project to Invoice / AP validation

  2. 02 · Data scraping · Flask web app · 2024–present

    Recall campaign site + VIN scraper

    Client project · freelance

    Problem Parse VIN-specific recall data from automotive sites.

    What it does · what I built A Flask web app with automated data scraping, CSV generation, anti-captcha solving, and background scheduling using Celery and Redis, with an admin dashboard for script monitoring and historical record management.

    Result Scraped over 1,000,000 records of VIN-specific recall data.

    How it works · step by step

    Client project · parts from the project notes · their dependency order reconstructed

    Input

    Automotive recall sites

    Where the VIN-specific recall data is published.

    Step · Celery · Redis

    Celery + Redis

    Background jobs schedule the scraping runs.

    Step · Python · Selenium · Anti-Captcha

    Scraping with anti-captcha

    Collects the recall data per VIN; captchas are solved during the run.

    Output

    CSV generation

    Results exported as CSV files.

    Output · Flask

    Flask admin dashboard

    Script monitoring and historical record management.

    Result One Flask app runs the scraping in the background and keeps its history, with CSV output.

    More on this case: Recall campaign site + VIN scraper

    Constraints

    • Anti-captcha solving
    • Background scheduling

    Approach

    Chose: Background scheduling using Celery and Redis, with an admin dashboard for script monitoring and historical record management

    PythonFlaskCeleryRedisSeleniumAnti-CaptchaCSV

    In my career: Freelance, via UpworkDiscuss a similar project to Recall campaign site + VIN scraper

  3. 04 · Automation · own build · Own build

    Content studio + distribution

    Own build · own brand

    Problem Hosted render APIs blew the budget, and nothing may go public without an approval gate.

    What it does · what I built Two n8n workflows. Content Studio (39 nodes): Claude ideates and scripts, ElevenLabs voices with word timestamps, fal.ai Flux makes scene images, an FFmpeg command renders on a self-hosted service. Distribution (26 nodes): a daily 9am run lists the approved folder, rules pre-flag, Claude writes a caption per platform, and it posts to FB Reels and YouTube. Deployable 24/7 on Railway.

    Result Estimated running cost: Claude (pennies) + ElevenLabs (~$11) + fal.ai (a few $) + self-hosted FFmpeg ($0) ≈ $11–15 a month. In the 6-video test batch, 3 are postable (10 platform posts).

    How it works · step by step

    Own build · two n8n workflows · order from their docs

    Rule · JavaScript

    Topic queue · pre-flag

    Only ready topics that were not produced already and are not scheduled for later go on.

    AI · Claude API

    Claude: script and scene plan

    Title, description, narration and a plan of scenes.

    Step · ElevenLabs

    ElevenLabs voiceover

    Narration to voice, with word timings for captions.

    Step · fal.ai Flux

    fal.ai scene images

    One illustrated still per scene, in a fixed style.

    Step · FFmpeg · self-hosted

    FFmpeg render

    Submit, wait, poll until done (an error stops the run), then download the MP4.

    Review · Google Drive

    Drive “Approved” folder

    The handoff. Distribution treats being in this folder as the approval signal; videos can also be added by hand.

    Rule · JavaScript

    Pre-flag content

    Approved, not already posted, inside its schedule window, within each platform's length limit.

    Rule

    Held or skipped

    A video still pending review or scheduled ahead waits; a platform it is too long for is skipped.

    AI · Claude API

    Claude: one caption per platform

    Captions and hashtags per platform; writes a title when one is missing.

    Output · YouTube

    YouTube

    Upload.

    Output · Facebook

    Facebook Reels

    Reels upload.

    Output

    Sheets log · Slack summary

    Every post and its status, plus a run summary.

    Result A topic becomes a published video with captions per platform, and nothing posts unless it is in the Approved folder.

    More on this case: Content studio + distribution

    Constraints

    • Hosted render APIs were over budget
    • Approval gate before anything goes public
    • Per-platform length limits

    Key decision

    Chose: Self-hosted FFmpeg render sidecar

    over Creatomate

    because Cost, no watermark, unlimited minutes

    Planted edge cases in the test batch
    CaseInputExpected
    Clean, approvedApproved videoPosts
    Pending reviewVideo still pending_reviewHeld
    Already postedVideo already postedSkipped
    Over X limitCaption or video over the X limitThat platform skipped
    Missing titleNo title suppliedClaude writes one
    Scheduled aheadScheduled for a later timeHeld

    n8nClaude APIElevenLabsfal.ai FluxFFmpegRailwayYouTubeFB Reels

    Discuss a similar project to Content studio + distribution

Client feedback · Upwork

What clients say

Reviews from Upwork contracts, quoted as written. Titles and dates are the contracts’ own.

  1. 5.0 / 5 rating

    “Recommended to everyone, the instructions, explanations and directions are very clear. 5 stars! for the service. I'll rehire this person again for our future projects.”

    Domain Migration with web hosting test and Updates. Fetch source code in name cheap

    Upwork client

    • Clear Communicator
    • Accountable for Outcomes
    • Detail Oriented
  2. 5.0 / 5 rating

    “Allan has been a pleasure to work with! The quality of his work as been great and his communication is clear.”

    Dev/Tech All-Rounder with Claude/AI Vibe Coding Enthusiasm to Solve Problems!

    Upwork client

    • Committed to Quality
    • Solution Oriented
    • Clear Communicator
  3. 5.0 / 5 rating

    “I've used Allan several times now and he's always been professional, on-time and gives me excellent service. I highly recommend him.”

    clean up

    Upwork client

    • Reliable
    • Professional
  4. 5.0 / 5 rating

    “Allan works has great quality results in our project and clear detailed communication. I recommend this to everyone 5 stars!! I will rehire him again for our project.”

    E-commerce Site Minor Fixes

    Upwork client

    • Committed to Quality
    • Clear Communicator
    • Detail Oriented
    • Accountable for Outcomes
  5. 5.0 / 5 rating

    “As always...A1 service! Corrected the issues perfectly and on time! I wouldn't use anybody else.”

    fix the password

    Upwork client

    • Reliable
    • Committed to Quality
    • Solution Oriented
    • Accountable for Outcomes
    • Detail Oriented
  6. 5.0 / 5 rating

    “Allan, gave the best service on Upwork and any service provider that I have received. He really took care of everything from A-Z. I couldn't be more happier. Great quality...”

    Excerpt as supplied

    Reddit Scraper for Market Research: Export to Google Docs

    Upwork client

    • Committed to Quality
    • Detail Oriented
    • Accountable for Outcomes
  7. 5.0 / 5 rating

    “Working with Allan was an absolute pleasure. They were professional, reliable, and delivered high-quality work ahead of schedule. Communication was smooth throughout the project, and they fully understood the requirements...”

    Excerpt as supplied

    Python API hosting

    Upwork client

    • Clear Communicator
    • Collaborative
    • Committed to Quality
    • Detail Oriented
    • Professional
    • Reliable
    • Solution Oriented
  8. 5.0 / 5 rating

    “Allan was very professional, ahead of schedule and I only had to tell him once what I needed. I highly recommend him for any of your Python or Excel related...”

    Excerpt as supplied

    MLS Script - Modification and Excel Cleanup

    Upwork client

    • Reliable
    • Solution Oriented
    • Clear Communicator
    • Accountable for Outcomes
    • Professional
  9. 5.0 / 5 rating

    “Allan is a very communicative person with good english. He is skilled and experienced in his area of expertise. Not only did he always answer to all my questions but...”

    Excerpt as supplied

    Ci/CD Workflows, Concept and Practical Integration

    Upwork client

    • Clear Communicator
    • Collaborative
  10. 5.0 / 5 rating

    “excellent work”

    Python Coder for Small Projects

    Upwork client

    • Committed to Quality
  11. 4.8 / 5 rating

    Rating only · no written feedback

    Pines Federal

    Upwork client

  12. 5.0 / 5 rating

    Rating only · no written feedback

    Fix both the MLS and Pacer passwords

    Upwork client

Showing reviews 1 of 12

See the Upwork profile (opens in a new tab)Selection supplied by Allan from his Upwork feedback; not live-synced.

How the work is put together

Capabilities

  • Data extraction and processing
  • Full-stack business applications
  • Workflow and AI automation
  • Deployment, maintenance and support

Combined differently in each project, not a fixed pipeline. Choose one to see an example and a project.

01 · extract

Extract

What it’s for Get the data you need out of websites, PDFs and APIs.

Python scrapers against static and dynamic sites, with proxy rotation and CAPTCHA handling.

  • Python
  • Scrapy
  • Selenium
  • Beautiful Soup
  • Requests
  • APIs
  • Proxy Rotation
  • Anti-Captcha
See it in a project: Legal case PDFs

02 · transform

Transform

What it’s for Turn raw, uneven records into clean tables you can use.

Cleaning, structuring and exporting to CSV or a database.

  • Pandas
  • NumPy
  • Data Cleaning
  • CSV/JSON Processing
  • PostgreSQL
  • MySQL
See it in a project: Shooju Global: clean CSV or database output

03 · automate

Automate

What it’s for Let rules and AI handle the routine cases and route the rest to a person.

n8n workflows where rules detect and Claude or OpenAI communicates, with background queues and schedules.

  • n8n
  • Claude API
  • OpenAI API
  • Celery + Redis
  • Background Queues
  • ElevenLabs
  • FFmpeg
  • Scheduled Workflows
See it in a project: Invoice / AP validation

04 · build

Build

What it’s for Web applications for the records and processes a team works with.

Full-stack web applications with Flask, Django and React.

  • Django
  • Flask
  • React
  • Next.js
  • TypeScript
  • Tailwind CSS
  • REST APIs
See it in a project: PAGASA: ICT Management System
ICT Management System: part of its maintenance table

Actual screen · ICT Management System (Django)

05 · ship

Ship

What it’s for Deploy it, automate the releases, and keep it running.

Docker, GitHub Actions and Ubuntu deployments; Railway for the n8n services.

  • Docker
  • Railway
  • Ubuntu Server
  • Nginx
  • GitHub Actions
  • Vercel
  • Google Cloud Storage
See it in a project: CI/CD project, on the map

Contact

What can I help you automate?

Based in the Philippines. Working with clients worldwide.

I'm open to freelance work, collaborations, and consulting opportunities. If you have a project in mind or just want to say hello, feel free to reach out.

  1. Book a free call, or email me. Tell me what you want to automate or improve, the tools you currently use, and where the process gets stuck.
  2. I reply with a plan. The approach I'd take, what I'd build, and a quote.
  3. I build, test and hand over. You review it working before it's done.

Book a free consultation20 minutes · online · no obligationEmail meallanboiebiaspunzalan@gmail.com

Remote work so far: Shooju Global (headquarters in Houston) and freelance clients in the United States, Australia, Guam, the United Kingdom and Europe.