Experience & Projects
No projects match your search
Try clearing search text or resetting active filter tags to discover more.
NetAiCore: Affinity Network & NLP Analysis Ecosystem
NetAiCore: Affinity Network & NLP Analysis Ecosystem
Designed a 360-degree analytics ecosystem and secure anonymization pipeline for chatbot operational databases, extracting sociodemographic insights under LFPDPPP privacy regulations.
Key Achievements
-
Built an advanced NLP pipeline utilizing spaCy Spanish models and BERTopic for sentiment analysis and topic modeling on conversational logs.
-
Developed a dual-mode Streamlit dashboard (internal with PII / shareable anonymized) and Quarto reporting scripts for stakeholders.
-
Modelled user interactions as a social graph with NetworkX and grouped user cohorts using HDBSCAN and scikit-learn clustering.
Technology Stack
Radar Comercial Querétaro V2: Geospatial License Analytics Dashboard
Radar Comercial Querétaro V2: Geospatial License Analytics Dashboard
Built a high-performance geospatial analytics dashboard to visualize over 110,000 active and blocked commercial business licenses in Querétaro, mapping Redshift datasets with interactive bounding boxes and metadata export pipelines.
Key Achievements
-
Engineered an interactive MapLibre GL JS mapping interface showcasing spatial business distributions and density filters across Querétaro.
-
Integrated dual-source data fetching connecting a production Amazon Redshift data warehouse with an optimized local CSV fallback parser.
-
Designed a sleek, modern UI utilizing Next.js 15, Tailwind CSS, and shadcn/ui components for rapid bounding-box and metadata exports.
Technology Stack
BrokerAI: Conversational AI & Lead Qualification Engine
BrokerAI: Conversational AI & Lead Qualification Engine
Architected and built a production-grade WhatsApp conversational sales agent and real-time operator dashboard for progressive lead qualification and automated appointment booking.
Key Achievements
-
Designed a robust AI orchestrator using OpenAI GPT-4o and WhatsApp Cloud API for multi-stage lead qualification and dynamic booking slots via Google Calendar API.
-
Developed a real-time operator workspace using Next.js 15, React 19, Tailwind CSS 4, and Server-Sent Events (SSE) for seamless human takeover.
-
Built a secure database schema with Prisma ORM on Cloud SQL (PostgreSQL), deploying cron jobs for appointment reminders and SLA escalation alerts to Slack.
Technology Stack
Meridian: Low-Code Territory Intelligence SaaS
Meridian: Low-Code Territory Intelligence SaaS
Architected a low-code SaaS platform mapping BigQuery-scale Mexican socio-demographic and geographic data into dynamically generated, explainable geospatial applications.
Key Achievements
-
Built a multi-agent orchestrator in LangGraph with 7 specialized agents (Profiler, Discovery, SQLGenerator, etc.) to translate natural language queries into verified SQL.
-
Integrated a large-scale data lake of ~577M rows (~79GB) on Google BigQuery using strict dry-run cost gates and topological dictionaries to protect query costs.
-
Designed IaC with Terraform for scalable deployment on GCP Cloud Run and integrated design tokens 1:1 using the BlackPrint Design System.
Technology Stack
SJS Sales Intelligence: Geospatial Predictive Modeling
SJS Sales Intelligence: Geospatial Predictive Modeling
Developed a predictive sales intelligence system and interactive GIS simulator to model branch performance and analyze competitive attraction in Querétaro.
Key Achievements
-
Designed a machine learning pipeline using XGBoost and ElasticNet to predict sales based on spatial competition, urban attractors, and mobility parameters.
-
Built a high-fidelity React frontend with MapLibre GL JS to display interactive target markets, custom radar charts, and multi-scenario side-by-side comparisons.
-
Implemented a FastAPI server delivering fast, on-demand inference for new-branch placement simulations with real-time UI updates.
Technology Stack
Proxy Security Risk Index: Neighborhood Safety Modeling
Proxy Security Risk Index: Neighborhood Safety Modeling
Designed a multi-factor geospatial proxy security index at the neighborhood (colonia) level for Ecatepec, modeling spatial crime correlation and public exposure metrics.
Key Achievements
-
Engineered a GeoPandas ETL pipeline integrating population census data (INEGI), business directories (DENUE), major transit corridors, and municipal crime figures (SESNSP).
-
Calculated normalized per-capita exposure indices using WGS 84 / UTM Zone 14N projections and classified risk levels via Jenks Natural Breaks.
-
Conducted advanced spatial autocorrelation using Local Moran's I (LISA) to detect crime spillover effects and hotspots near inter-municipal transit terminals.
Technology Stack
Blackprint Radar: Multi-Source POI Matching Pipeline
Blackprint Radar: Multi-Source POI Matching Pipeline
Engineered a high-performance, multi-stage deduplication and matching pipeline to merge, reconcile, and spatially enrich Point of Interest (POI) data from commercial and government sources in Querétaro.
Key Achievements
-
Developed an adaptive text-similarity algorithm combining Jaro-Winkler, Levenshtein, and semantic category similarity using paraphrase-multilingual sentence embeddings.
-
Implemented a high-performance SpatialGrid cell bucketing algorithm to optimize hot-loop deduplication matching across 215k x 34k records in under 3 minutes.
-
Engineered cadastral spatial joins using STRtree to enrich POIs with 2023 municipal zoning guidelines and verify business compatibility (AMAI / SCIAN classification).
Technology Stack
H Barbería Colima: Client Booking & Management Web App
H Barbería Colima: Client Booking & Management Web App
Designed and developed a client-facing booking web app for a local barbershop in Colima, integrating custom calendar scheduling parameters per-barber and modern fluid galleries with SEO optimization.
Key Achievements
-
Developed a fluid scheduling workflow using React 19 and Vite 6, incorporating Google Calendar sync integrations to eliminate double bookings.
-
Implemented responsive CSS modules with premium glassmorphism aesthetics and custom micro-animations to increase user engagement.
-
Optimized load times and SEO rankings using React Helmet Async and semantic HTML, ensuring zero-latency transitions.
Technology Stack
OCR Gemini: Certified Legal Translation Pipeline
OCR Gemini: Certified Legal Translation Pipeline
Led and guided a team of interns to develop a high-fidelity Python CLI pipeline using Gemini Vision API, instructing them on OCR best practices and spatial layout mapping to automate certified translations of multi-page legal documents.
Key Achievements
-
Mentored interns on implementing a vision-based OCR pipeline utilizing Gemini Vision to extract and translate low-resolution legal scans with unmatched accuracy.
-
Guided the team in building a layout-preserving DOCX compiler that maps text positions, tables, and headers into a professional bilingual format.
-
Instructed junior engineers on integrating automatic media resolution preprocessing, reducing token costs while optimizing detection rates for certified signatures.
Technology Stack
Meeting Intelligence Platform
Meeting Intelligence Platform for Legal Teams
Led the development and guided a team of interns to build a meeting intelligence system that converts Zoom/Teams/Meet recordings into searchable transcripts and structured summaries with group-based access control.
Key Achievements
-
Mentored interns in automating transcription and summarization workflows using Speechmatics and LLM APIs for legal case documentation.
-
Guided the team in building semantic search across meetings using Pinecone vector indexing and retrieval.
-
Instructed interns on implementing secure RBAC controls for admin, supervisor, and user access levels under industry best practices.
Technology Stack
DeclarationLetterOnline: Legal AI Generator
DeclarationLetterOnline: Legal AI Generator
Led and directed a team of interns to build a production-grade FastAPI and React legal document generator utilizing Gemini 1.5 Pro, instructing them on prompt engineering, strict document styling, and secure local audit trails.
Key Achievements
-
Guided interns on developing a multi-stage NLP synthesis pipeline that ingests questionnaires and translates them into certified legal letters in Century Schoolbook format.
-
Mentored the team in building an asynchronous processing workflow with local SQLite audit trails and structured XML parsing for automated DOCX compilation.
-
Instructed junior developers on creating a React interface featuring real-time feedback loops, version control of draft revisions, and instant side-by-side editing.
Technology Stack
LLM GeoAI Platform
LLM GeoAI Platform for Urban Intelligence
Built an LLM-powered GeoAI platform to model 2.5M+ urban blocks with real-time inference and interactive mapping for faster decision-making.
Key Achievements
-
Shipped React + Mapbox + Turf.js pipelines for scalable geospatial visualization.
-
Integrated LLM reasoning to surface location intelligence insights in production.
-
Delivered real-time analytics to support licensing and urban planning workflows.
Technology Stack
Electoral and Socioeconomic Geospatial Analysis
Electoral and Socioeconomic Geospatial Analysis
Developed an advanced analytical platform integrating historical electoral data (2018-2024) with socioeconomic indicators at the block level for all of Mexico, allowing multidimensional analysis of electoral behavior. Implemented a frontend with React 18, TypeScript 5.0, and Mapbox GL JS v2.15 with custom extensions for interactive geospatial visualization and spatial analysis using Turf.js.
Key Achievements
-
Designed and implemented Python scripts with GeoPandas for geospatial ETL, shapefile transformation, and vector processing for geometric schema unification.
-
Processed more than 2.5M urban blocks from INEGI with more than 30 variables per entity, implementing AMAI classification algorithms for socioeconomic level calculation.
-
Implemented topological optimization with adaptive Douglas-Peucker simplification and generation of vector tiling schemes with efficient encoding.
Technology Stack
Medical QA Multi-Agent System
Medical QA Multi-Agent System
Built a multi-agent LLM system grounded in PrimeKG and Neo4j to improve medical QA accuracy through multi-hop reasoning.
Key Achievements
-
Designed specialized agents for clinical reasoning and evidence retrieval.
-
Integrated knowledge graph queries to support explainable QA outputs.
-
Evaluated system performance with systematic benchmarking.
Technology Stack
Intelligent Commercial Document Analysis System
Intelligent Commercial Document Analysis System
Built an OCR and Llama 3.3 document extraction pipeline for invoices, packing lists, and customs documents, delivering structured outputs at scale.
Key Achievements
-
Deployed the extraction API with FastAPI and a React/Vite frontend for real-time PDF processing.
-
Implemented multi-format ingestion (PDF, CSV, XLSX) with asynchronous workers.
-
Improved extraction accuracy through targeted model tuning and validation.
Technology Stack
Specialized Industrial Real Estate Chatbot
Specialized Industrial Real Estate Chatbot
Designed and implemented a custom RAG (Retrieval Augmented Generation) system with chatbot for the industrial real estate sector, integrating Anthropic's Claude 3.5 Sonnet and OpenAI APIs. Developed a data processing system for semantic vectorization (embeddings) of complex real estate information using OpenAI Embeddings and FAISS for efficient search.
Key Achievements
-
Implemented advanced prompt engineering techniques with a multi-layer system for contextualization, generation, and formatting, significantly improving the relevance and accuracy of responses.
-
Created a dynamic PDF report generation system from conversations, using Jinja2 for templating and HTML-PDF conversion techniques for structured documents.
-
Developed a customized user interface in Streamlit with multi-user authentication system, dynamic theming, and accessibility optimizations.
Technology Stack
Comprehensive Data Analysis Platform
Comprehensive Data Analysis Platform for TV Shopping Company
Developed a complete data analysis and business visualization ecosystem consisting of three interconnected projects: a RESTful API backend (FastAPI), an interactive dashboard (React), and a report generation platform (Flask with AI). Implemented a RESTful API with FastAPI that provides access to analytical data stored in MongoDB, with endpoints for different dimensions of business analysis.
Key Achievements
-
Designed and developed an interactive dashboard with React 18 and Material UI 6 for business KPI visualization, implementing visualizations with Recharts, D3.js, and React Leaflet.
-
Created a data-driven decision-making solution, implementing KPIs for sales trends, customer segmentation, inventory turnover, and advertising effectiveness through interactive dashboards.
-
Employed K-means clustering for customer segmentation and created interactive chloropleth maps using GeoJSON data, revealing geographical patterns in sales and customer behavior.
-
Integrated Google Trends data using time series analysis to anticipate consumer interests, informing inventory management and marketing strategies.
-
Implemented a multi-stage data processing pipeline, extracting data from SQL databases, storing in MongoDB, processing with pandas and numpy, and integrating external APIs for data enrichment.
Technology Stack
Advanced Demand Forecasting & Inventory Optimization System
Advanced Demand Forecasting & Inventory Optimization System
Developed a complete demand forecasting and inventory optimization system for the beverage industry, incorporating time series analysis, statistical modeling, and full-stack implementation. Designed and built a robust ETL pipeline for beverage sales data, with multiple source integration, advanced cleaning with pattern extraction, and automated outlier detection.
Key Achievements
-
Implemented an adaptive forecasting system that selects optimal algorithms based on data characteristics: Prophet for SKUs with sufficient history, SARIMA for products with moderate history, and moving averages as fallback.
-
Developed a mathematical inventory optimization model incorporating lead time calculations, safety stock determination, capacity constraints, and turnover rate optimization.
-
Created a modern and responsive dashboard with React/TypeScript, interactive visualizations with Chart.js, component-based architecture, and performance optimization through memoization and lazy loading.
Technology Stack
GeoAI Insights: Advanced Visualization of the Competitive Landscape
GeoAI Insights: Advanced Visualization of the Competitive Landscape
Leveraged GPT-4 and GPT Vision for competitor analysis, using chain of thought reasoning and previous examples to extract insights on employee count, building size, and competitor locations. Implemented GIS tools with GeoJSON and ArcGIS Online, overcoming data cleaning challenges to create detailed maps of showroom and retail locations.
Key Achievements
-
Utilized Storymaps in ArcGIS Online for effective visualization, balancing local and cloud-based processing based on analysis requirements.
-
Developed gradient-based visualizations using ArcGIS API to highlight competition concentrations across the US, ensuring interpretability through interactive maps.
-
Integrated AI-driven analysis with geospatial data, validating accuracy with GPT Vision API, to create a comprehensive competitive landscape view.
Technology Stack
Interactive Personal Portfolio Development
Interactive Personal Portfolio Development
Designed and developed an interactive personal portfolio using Astro, React, and Tailwind CSS, optimizing for performance and SEO while leveraging popular frameworks for UI development. Implemented a responsive and accessible design following W3C Web Content Accessibility Guidelines (WCAG), ensuring optimal user experience across all devices and tested using browser developer tools.
Key Achievements
-
Created reusable components for clean and maintainable code, including navigation menu, footer, and contact form, which facilitate effective data visualization and user interaction.
-
Integrated smooth animations using Animate.css library to enhance interactivity and visual appeal, contributing to effective data storytelling and user engagement.
-
Optimized website performance using Lighthouse, Google PageSpeed Insights, and WebPageTest tools, achieving fast loading times and a fluid user experience with measurable improvements.
Technology Stack
Automated Data Analysis System
Automated Data Analysis System for Motor Vehicle Company
Developed an automated system for efficient analysis of large-scale web analytics data using Google Analytics metrics and dimensions, with on-demand data extraction and batch processing. Designed and implemented a comprehensive ETL pipeline, incorporating data cleaning, validation, and profiling, along with AI-driven insight generation using Large Language Models.
Key Achievements
-
Created an intuitive dashboard with filters and KPIs using HTML, CSS, and JavaScript, enabling non-technical stakeholders to select data and view reports easily.
-
Conducted country-specific analysis of web data, analyzing user behavior patterns and cultural nuances across metrics like Active Users, Engagement Rate, and device preferences, as well as country-specific KPIs such as Build & Price completions and test drive bookings.
-
Established thorough documentation practices using Google Docs and version control, enhancing project transparency and replicability.
Technology Stack
Comprehensive Time Series Analysis and Forecasting Project
Comprehensive Time Series Analysis and Forecasting Project
Developed a robust time series analysis and forecasting pipeline using Python, incorporating multiple advanced models including SARIMA, Prophet, XGBoost, LSTM, and Transformer. Implemented data preprocessing techniques such as handling missing values, resampling, and feature engineering to enhance model performance.
Key Achievements
-
Utilized concurrent processing with ThreadPoolExecutor to efficiently handle multiple SKUs, optimizing computational resources and reducing execution time.
-
Integrated advanced visualization techniques using Matplotlib and Plotly to create interactive and informative charts for trend analysis and forecast comparison.
-
Implemented inventory optimization algorithms, including EOQ and safety stock calculations, considering sustainability factors like CO2 emissions.
-
Leveraged GPU acceleration for deep learning models (LSTM, Transformer) using PyTorch and TensorFlow, significantly improving training speed.
-
Developed a custom Transformer model architecture for time series forecasting, showcasing adaptability to complex sequential data.
-
Implemented robust error handling and logging mechanisms to ensure reliable execution across large datasets.
-
Utilized Optuna for hyperparameter optimization, enhancing model performance through automated tuning.
-
Integrated geospatial analysis using Folium to visualize geographical patterns in sales and inventory distribution.
Technology Stack
Automated Interview Analysis
Automated Interview Analysis
Automated survey data analysis from SurveyMonkey, reducing a 1-2 month process to seconds of script runtime. Implemented data normalization and weighted sampling to mitigate demographic biases in survey responses.
Key Achievements
-
Utilized sentiment analysis and topic modeling on open-ended responses to extract key themes and identify market opportunities.
-
Leveraged Python, pandas, and multi-index for complex data manipulations, creating multi-level pivot tables for cross-sectional analysis.
-
Optimized code performance using vectorized operations and chunking techniques for efficient large dataset processing.
-
Generated custom radar and multi-line charts with Plotly, embedded in HTML, enhancing data interpretability.
-
Balanced company design specifications with data visualization best practices, ensuring accessibility and effective communication.
-
Applied ANOVA and chi-square tests to identify statistically significant performance issues across groups and sections.
-
Developed an interactive HTML dashboard for management, featuring dynamic data exploration and highlighting key insights.
Technology Stack
Trading Strategy Backtesting
Trading Strategy Backtesting
Extracted financial data via Alpha Vantage API and visualized trading patterns using Plotly in Python, implementing rate limiting and error handling for data consistency. Configured a MongoDB database with a document-based schema for storing varied economic data, optimizing for scalability and query performance.
Key Achievements
-
Conducted correlation analysis on key economic indicators (GDP, unemployment, inflation, interest rates) versus market trends, implementing lag analysis for optimal impact assessment.
-
Discovered through rigorous testing that strategies performed poorly, using metrics like Sharpe ratio and maximum drawdown, highlighting the need for advanced techniques in future projects.
Technology Stack
Geospatial Analysis of Ciudad Guzmán
Geospatial Analysis of Ciudad Guzmán
Conducted exploratory geospatial analysis in ArcGIS using data from the Jalisco Seismic Network, utilizing time slider for temporal visualization. Analyzed patterns between earthquake intensity and elevation using Getis-Ord Gi* spatial statistics and Geographical Weighted Regression.
Key Achievements
-
Created an interactive Storymaps map in ArcGIS, using templates and pop-ups to present complex geospatial data accessibly.
Technology Stack
Data Analysis
Data Analysis for Property Management System
Delivered property analytics with Google Maps geocoding and robust pandas/SQL ETL pipelines for high-confidence operational reporting.
Key Achievements
-
Achieved 92% geocoding accuracy through address normalization and validation.
-
Built pandas/SQL ETL pipelines that ensured 99.9% deliverable accuracy.
Technology Stack
User Engagement Analysis
User Engagement Analysis for SaaS Platform
Analyzed user engagement using metrics like DAU, MAU, and feature adoption rates, segmenting users by subscription tier and usage frequency. Developed custom JavaScript visualizations in Looker Studio, creating interactive sankey diagrams and heatmaps.
Key Achievements
-
Translated analytical findings into strategic recommendations, creating layered dashboards in Looker Studio for different levels of technical understanding.