From idea to production: AI agents, secure systems, APIs and infrastructure.

Context

Product built for candidates who need to monitor multiple official sources without manually checking each administration website.

Role

Designed and built the scraping, data processing and notification workflows.

Problem

Information about public exams and interim opportunities is distributed across many administration websites, making it easy to miss relevant updates.

Solution

Built a system that monitors selected public sources, normalizes detected opportunities and sends targeted alerts to subscribed users.

Contribution

• Implemented scraping jobs for multiple public administration websites.
• Modeled opportunities, subscriptions and notification rules.
• Built the delivery workflow for notifying users when relevant updates appear.

Technologies

Outcome

• Centralized monitoring of fragmented public sources.
• Timely alerts for subscribers based on their interests.

Context

SaaS product for clinics that want a lightweight receptionist layer without adding more manual admin work.

Role

Designed the product workflow, WhatsApp conversation flow and calendar integration.

Problem

Clinics lose time answering repetitive scheduling messages and manually coordinating availability, confirmations and calendar updates.

Solution

Built a WhatsApp-based receptionist that handles appointment requests, checks availability and creates calendar events for subscribed clinics.

Contribution

• Modeled clinic onboarding, subscription state and appointment flows.
• Integrated WhatsApp messaging with automated scheduling logic.
• Connected bookings with Google Calendar to keep clinic calendars in sync.

Technologies

Outcome

• Reduced repetitive receptionist work for clinics.
• Automated appointment flow from message to calendar booking.

Context

Client project focused on turning fragmented review data into business insight and market signals.

Role

Building the data aggregation and analysis layer with a collaborator.

Problem

Businesses have access to many public opinions, but extracting trends, sentiment and competitive signals from them is time-consuming.

Solution

Aggregates opinions from different sources and applies a data analysis layer to surface sentiment, market positioning and recurring patterns.

Contribution

• Designed the data model for review aggregation and analysis.
• Built ingestion workflows for external opinion sources.
• Helped shape the insight layer for market and sentiment reporting.

Technologies

Outcome

• Transforms raw opinions into usable market intelligence.
• Helps businesses identify patterns that are difficult to see manually.

Context

Company project for scheduling recurring communications with configurable topics, audiences, frequency and delivery media.

Role

Built services for configuring, generating and delivering recurring AI-assisted reminders.

Problem

Teams needed a flexible way to remind different user groups about recurring topics without manually writing and sending every message.

Solution

Created a reminder system where teams can choose audience, topic, frequency and channel, with AI helping generate the message content.

Contribution

• Modeled audience selection, reminder frequency and campaign configuration.
• Built services for AI-assisted content generation and delivery scheduling.
• Integrated multiple notification channels for operational use.

Technologies

Outcome

• Automated recurring communication workflows.
• Gave teams control over audience, channel, topic and frequency.

Context

Company project for construction teams that report issues during the week using quick audio messages.

Role

Built the automation that processes incoming audio notes and produces weekly email drafts.

Problem

Users captured project problems as audio notes, but turning those notes into structured weekly communication required manual effort.

Solution

Built a bot that collects audio messages, processes them and generates a weekly email summary that users can review and send.

Contribution

• Designed the audio intake and weekly batching workflow.
• Implemented AI summarization for construction-related project notes.
• Generated email-ready summaries for end users.

Technologies

Outcome

• Reduced manual reporting effort for construction project updates.
• Converted unstructured audio notes into actionable weekly summaries.

Context

Company project built with teammates to support construction budgeting workflows with domain-specific knowledge.

Role

Contributed to the AI and data layer for budget generation using LLMs and retrieval over training data.

Problem

Creating construction budgets requires domain knowledge, structured cost data and repetitive manual work.

Solution

Built an AI-assisted budget generator that retrieves relevant construction data and uses LLMs to help produce structured BC3 estimates.

Contribution

• Worked on retrieval workflows over construction cost data.
• Helped integrate LLM output into budget generation processes.
• Collaborated on a production-oriented tool with domain-specific constraints.

Technologies

Outcome

• Accelerated budget drafting for construction workflows.
• Reused domain data to produce more contextual estimates.