Overview:

• Designed and developed the frontend for a Latin American iGaming platform, delivering a responsive, dark-themed SPA focused on performance, usability, and seamless integration with backend gaming services.
• The application provides a complete player-facing experience across game discovery, catalog navigation, authenticated account functionality, game sessions, localization, notifications, and real-time platform updates.

Technical Stack:
React, Vite, Redux Toolkit, React Router, REST APIs, Axios, Tailwind CSS, i18n, Socket.IO

Gaming Lobby & Catalog Experience:

• Built the main gaming lobby with hero sections, promotional cards, category navigation, provider filters, and dynamic game catalogs.
• Developed dedicated catalog experiences for casino, live casino, crash/arcade, and themed game collections.
• Implemented provider-based filtering and search functionality for quickly navigating large game catalogs.
• Integrated live prize and winners information into the lobby experience.
• Created reusable game-card and catalog components to maintain consistent UX across multiple sections.

Frontend Architecture:
Designed a feature-oriented frontend architecture using reusable application modules and custom hooks.
Key architectural areas included:

• User and authentication state
• Game catalogs and game sessions
• Transactions and account data
• Notifications
• Promotional banners
• Localization
• Shared layout and game data
  Used Redux Toolkit and React Context strategically to manage global application state while keeping feature-specific logic isolated and maintainable.

UI & User Experience:

• Built a responsive dark-themed interface optimized for an immersive gaming environment.
• Used Tailwind CSS and Material UI for reusable UI components and responsive layouts.
• Implemented animated interactions using Framer Motion.
• Created reusable navigation, filtering, modal, catalog, notification, and account components.
• Focused on responsive behavior across desktop and mobile screen sizes.

Outcome:

• Delivered a production-ready frontend foundation for a multilingual iGaming platform, supporting dynamic game catalogs, provider filtering, authenticated user experiences, full-screen game sessions, account management, real-time updates, and responsive interfaces.
• The frontend architecture was designed to support the continued expansion of game providers, catalogs, user features, localized content, and backend integrations.

Overview:

• Designed and developed a full-stack tire pricing and product intelligence platform for a Spanish tire retailer, automating the collection, normalization, monitoring, and analysis of thousands of tire SKUs.
• The platform continuously tracks product specifications, prices, stock availability, EU labels, images, and product information, allowing the retailer to monitor market pricing, identify price reductions, and maintain an up-to-date product catalog with minimal manual effort.

Technical Stack:
React, FastAPI, Selenum, SQLAlchemy, REST APIs, Vite, BeautifulSoup, PostgreSQL

Automated Data Collection & Web Scraping:

• Built a robust Python web-scraping pipeline using BeautifulSoup and Cloudscraper for automated product data extraction.
• Extracted and normalized tire information including product names, brands, specifications, prices, stock levels, EU labels, images, and product URLs.
• Implemented handling for Spanish pricing and number formats to ensure accurate storage and comparison of product prices.
• Added request delays, retries, and error handling to improve scraper reliability and resilience.
• Designed the scraping workflow to support both targeted product refreshes and large-scale catalog updates.
• Structured scraped data for reliable downstream processing, filtering, reporting, and price analysis.

Architecture & Engineering:
The platform was designed around an automated pipeline:
Web Sources → Scraping → Data Normalization → PostgreSQL → Price Analysis → FastAPI → Administration Dashboard

Outcome:

• Delivered an automated tire pricing and product intelligence platform that replaced manual product monitoring with a centralized system for collecting, processing, analyzing, and exporting large-scale tire catalog data.
• The platform provides the retailer with continuous product-data refreshes, historical price tracking, automated price-drop detection, operational monitoring, and scalable catalog management, creating a strong foundation for further expansion into competitive pricing intelligence, automated market monitoring, and AI-powered product analysis.

Overview:

• Designed and developed an AI-powered retail intelligence SaaS platform that transforms retail and supermarket data into actionable business insights.
• The platform combines structured retail data, data-processing pipelines, and LLM-powered analysis to help turn complex business information into understandable insights that can support operational and strategic decision-making.
• Worked across the complete application stack—from the Next.js frontend and FastAPI backend to PostgreSQL, data processing, and AI/LLM integration.

Technical Stack:
Next.JS, Python, FastAPI, REST APIs, PostgreSQL, AI Pipelines, API Integrations, Data Processing

AI & Retail Intelligence:

• Integrated LLM capabilities into the retail analytics workflow to transform structured business data into contextual, actionable insights.
• Designed AI workflows that combine retail data, business context, and structured prompts to produce relevant results rather than generic AI responses.
• Prepared and transformed application data into structured model inputs suitable for LLM analysis.
• Processed and validated model responses before returning AI-generated insights through the application.
• Designed the AI layer to work as part of the existing business workflow rather than as a standalone chatbot.
• Created a foundation for expanding the platform with additional AI-powered retail analytics and decision-support capabilities.

Retail Data Processing:

• Built data-processing logic for converting raw and structured retail information into AI-ready datasets and contextual inputs.
• Worked with retail and supermarket business data across multiple application workflows.
  Structured contextual information so AI analysis remained grounded in the underlying business data.
• Connected processed retail data with AI analysis workflows to generate meaningful, data-driven outputs.
  Designed processing flows with scalability and maintainability in mind as the volume and complexity of retail data grows.

AI Application Architecture:
Designed the application flow around a clear separation between:
Retail Data → Data Processing → Context & Prompt Construction → LLM Analysis → Response Processing → API → AI Insights

Outcome:

• Delivered a production-oriented AI retail intelligence platform that connects real retail business data with LLM-powered analysis, enabling users to transform complex datasets into contextual and actionable insights.
• The architecture provides a strong foundation for expanding into AI-powered retail forecasting, automated reporting, anomaly detection, business intelligence assistants, and intelligent decision-support workflows.

Overview:
Designed and developed a full-stack AI-powered career assessment and job recommendation platform that helps users identify suitable career paths through a personalized assessment.
The platform combines a deterministic scoring and ranking engine with LLM-powered explanations to analyze user interests, match them against a structured catalog of 5K+ job roles, and deliver personalized career recommendations with clear, human-readable reasoning.

Technical Stack:
Next.JS, Python, FastAPI, React, TypeScript, LLM, REST APIs, CI/CD, PostgreSQL

What I Delivered:

1. Intelligent Career & Recommendation Engine

• Built an end-to-end career assessment workflow covering questionnaire completion, email verification, scoring, and personalized recommendations.
• Developed a configurable recommendation engine supporting 5K+ job roles across multiple career sectors.
• Designed a deterministic scoring and ranking system to ensure consistent, explainable recommendations.
• Combined rule-based ranking with AI-generated insights to provide personalized explanations for recommended career paths.
• Implemented sector-aware recommendation logic to maintain diversity across career fields rather than returning overly similar roles.
• Added configurable scoring rules, thresholds, ranking logic, and sector preferences to allow recommendation behavior to evolve without rewriting the core system.
• Built structured recommendation results containing job titles, sectors, match information, and personalized explanations.

2. AI & LLM Integration

• Integrated Google Gemini to generate natural-language explanations based on structured assessment and recommendation data.
• Designed the AI layer to complement deterministic recommendation logic rather than allowing the LLM to independently determine rankings.
• Generated personalized career insights that explain why a specific role is a strong match for each user.
• Structured AI inputs and outputs to maintain consistency and reliability across recommendations.

3. Automation & Integrations

• Integrated the Systeme.io API for automated lead capture and marketing workflows.
• Implemented automated contact creation, tagging, and email workflow triggers.
• Designed the integration layer to support additional third-party services as the platform expands.

The architecture provides a strong foundation for expanding into a larger AI-powered career discovery platform, with opportunities for advanced personalization, analytics, career planning, learning recommendations, and additional AI capabilities.

Overview:

• Designed and developed a full-stack AI-powered Forex trading and operations platform integrated with MT4, MT5.
• The platform combines AI-assisted market analysis, algorithmic trading workflows, risk management, automated execution, backtesting, and real-time monitoring into a unified system.
• The architecture separates AI analysis from deterministic trading and risk controls, allowing AI-generated insights to support decision-making while predefined rules remain responsible for execution and operational safeguards.

Technical Stack:
AI/ML, Next.JS, Python, FastAPI, PostgreSQL, WebSocket APIs, Docker, CI/CD

AI-Powered Trading Intelligence:

• Integrated AI capabilities to analyze structured market and trading data and generate actionable trading insights.
• Built an AI-assisted analysis layer for identifying market patterns, signals, trends, and contextual trading conditions.
• Used LLM-based reasoning to transform structured trading information into human-readable market analysis.
• Designed AI outputs to work alongside deterministic strategy and risk rules rather than directly bypassing execution safeguards.
• Created a foundation for AI-assisted strategy analysis, trade explanations, and performance insights.
  Structured the AI layer so additional models and analytical workflows can be integrated as the platform evolves.

Outcome:

• Delivered a full-stack AI-powered trading platform that combines AI-assisted market intelligence with algorithmic execution, configurable risk management, backtesting, and real-time operational monitoring.
• The architecture provides a strong foundation for expanding into more advanced AI trading agents, predictive analytics, strategy optimization, automated performance analysis, and intelligent trading assistants while maintaining deterministic controls around execution and risk.
• Trading profitability is dependent on strategy quality, market conditions, execution, and rigorous validation. The platform is designed to provide the technology infrastructure and controls for systematic, AI-assisted trading—not to guarantee financial returns.

Built a scalable backend platform for a US company directory containing approximately 100,000 company records, including company profiles, locations, industries, websites, and executive/CEO information.

Key Contributions:

• Designed and developed the backend architecture using Python and FastAPI, providing reliable REST APIs for searching, filtering, retrieving, and managing company and executive data.
• Designed a structured PostgreSQL data model for companies, executives, and related attributes, including relationships and indexes optimized for high-frequency queries.
• Built API endpoints for company search, advanced filtering, pagination, company profiles, and executive information.
• Implemented optimized company and executive name search using PostgreSQL indexing and trigram-based text search, improving fuzzy matching and search performance across large datasets.
• Optimized database queries and API responses to minimize unnecessary data retrieval and maintain responsive performance as the dataset grew.
• Architected the search layer with Elasticsearch in mind for more advanced full-text search, fuzzy matching, relevance scoring, and future scalability.
• Implemented Pydantic validation and structured API responses to ensure consistent, predictable data contracts for frontend and third-party integrations.
• Worked with large structured datasets and backend performance optimization, focusing on scalability, search accuracy, maintainability, and production readiness.

Result:
Delivered a production-ready backend foundation for a large-scale company directory, capable of efficiently serving structured company and executive data through well-defined APIs while providing a scalable path toward advanced search capabilities.

Skills:
Python, FastAPI, PostgreSQL, REST APIs, Elasticsearch, Data Search & Optimization