Skip to content
ContactChat on WhatsApp(opens WhatsApp in a new tab)
Português

Portfolio

Projects

Work described anonymously and grouped by type of problem: the context, what was done, the outcome and the technologies. No client names and no implementation details. The only demo model is labeled as such.

  • AI communication hub with human approval

    Context
    Conversations, requests and agreements from an entire messaging channel were stuck in the chat history, with no triage, no record and no way to find the context of an old decision.
    Approach
    Messages collected into a dedicated store, contextual history, triage with filters and tags, search for tasks and pending items, AI-suggested replies and manual approval before anything is sent.
    Outcome
    Communication organized in a single dashboard, with context that can be found and nothing sent without human confirmation.
    See the layered architecture
    • Linux
    • Node.js
    • LLMs
    • Local APIs
    • Authentication
    • Logs

  • Capacity planning and schedule validation

    Context
    The operation needed to size its team in line with real demand, without relying only on gut feeling or a simple average.
    Approach
    Analysis of months of volume by day and by hour, handling of atypical days, calculation of each slot’s weight, forecasting, sizing and schedule validation.
    Outcome
    Planning that can be justified and repeated, with forecast, capacity and schedule kept separate and the criteria documented.
    See how the problem shows up
    • Excel
    • Power BI
    • Python
    • Time series
    • Workforce management

  • Automatic cross-system checks with OCR

    Context
    Decisions and identifiers shown in different systems, with no integration between them, had to be compared quickly to catch discrepancies.
    Approach
    Screen content captured with OCR, identifiers normalized and compared against rules, and inconsistencies flagged while the work is happening.
    Outcome
    Faster verification and an objective layer of checking, with discrepancy types classified to cut false alarms.
    • OCR
    • Python
    • UI automation
    • Business rules

  • Analysis of overtime and payroll components

    Context
    The origin of a consolidated overtime total had to be explained, and the amounts then broken down by person.
    Approach
    The total rebuilt from the individual entries, each component validated and the amounts distributed by employee in order of impact.
    Outcome
    An aggregate number turned into an auditable analysis, traceable from the grand total to pay items and individuals.
    • Excel
    • Power Query
    • Python
    • Total reconciliation

  • Audit and reconciliation of people databases

    Context
    Two operational databases differed in names and in which records were present, and there was no way to confirm that everyone was on both sides.
    Approach
    A full comparison of the lists, data normalization, exact and fuzzy name matching, and an audit file as the output.
    Outcome
    Real differences separated from spelling variations, less manual checking and a reusable basis for future audits.
    • Excel
    • Python
    • pandas
    • openpyxl
    • Matching rules

  • Access and offboarding governance

    Context
    Accounts and permissions needed a review to reduce unnecessary access and operational risk.
    Approach
    An inventory of users, groups, permissions and usage, identification of inactive accounts and criteria for removal and review.
    Outcome
    Less exposure to stale access and a more controlled offboarding process, with the repetitive steps automated.
    • Active Directory
    • GPO
    • Scripts
    • Logs
    • Inventory
  • Demo model

    Executive reporting model

    Context
    A leadership report assembled by hand every month, with numbers of untraceable origin and deviations noticed only after the month had closed.
    Approach
    An Excel model with 9 source tables, an isolated calculation engine, 16 automatic alert rules and 11 reconciliation checks, feeding 12 report pages from a single control cell.
    Outcome
    No number typed in by hand: changing the reference month updates the entire report, and any leadership metric can be traced back to its source table.
    See the example dashboard
    • Advanced Excel
    • Named ranges
    • Driver-based forecasting
    • Total reconciliation

  • Automated radar for technical updates

    Context
    Following several technology areas by hand meant too many sources and the risk of missing relevant updates.
    Approach
    A pipeline for collection, deduplication, classification, embeddings and relevance scoring, with daily and weekly reports written by a local model.
    Outcome
    Scattered monitoring turned into a structured technical intelligence process, isolated and validated before running on its own.
    • Python
    • SQLite
    • RSS and APIs
    • Embeddings
    • Local model

  • Publishing with Git, a test environment and a custom domain

    Context
    A website had to be developed and published safely, keeping every version and with no direct changes in production.
    Approach
    A repository with branches, pull requests, a test environment, static publishing and domain setup through DNS.
    Outcome
    Development, validation and production kept separate, with a change history and the ability to roll back.
    • Git
    • GitHub
    • GitHub Pages
    • DNS

  • Technical SEO, semantics and accessibility audit

    Context
    A website had to be prepared for launch with better indexing, semantic structure and technical consistency.
    Approach
    Review of headings, metadata, robots.txt, sitemap, Search Console, indexability, alt text and semantic elements.
    Outcome
    Lower technical indexing risk, a better structure for search engines and assistive technologies, and a reusable launch checklist.
    • Technical SEO
    • Semantic HTML
    • Accessibility
    • Search Console

Contact

Have a problem like one of these?

The projects show the kind of problem, not the limit of what can be done. Describe what is holding things up.

Chat on WhatsApp(opens WhatsApp in a new tab)Prefer email? Other ways to get in touch