Daniel Mora

AI Engineer & Full-Stack Developer

I build AI products, full-stack systems, and practical tools that turn complex ideas into useful software.

About Me

Hey, I'm Daniel. I'm a software engineer focused on AI, automation, and full-stack products. I like building systems that are useful in the real world: chatbots, recommendation engines, cloud apps, data pipelines, client websites, and the occasional game when the creative itch wins.

IE University
AI & ML Builder
3 Languages
Daniel Mora - AI & Software Engineer

Featured Projects

Open the repos, model releases, and project details

Alien Invasion

Designed and implemented a 2D shooter game in C++, featuring real-time enemy AI, dynamic difficulty scaling, and engaging gameplay mechanics.

C++ SFML OpenGL SDL
Alien Invasion Gameplay Alien Invasion Enemies Alien Invasion Shooting Mechanics Alien Invasion Scoring System

Game Features

  • ๐ŸŽฎ Real-time Enemy AI
  • ๐Ÿ“ˆ Dynamic Difficulty Scaling
  • ๐ŸŽฏ Pathfinding AI
  • ๐Ÿ’ฅ Optimized Game Loop
  • ๐ŸŽจ 2D Graphics & Rendering

Technologies Used

  • C++: Object-oriented programming
  • SFML: Game development framework
  • OpenGL & SDL: Graphics & rendering
  • GitHub: Version control

AI Sales Chatbot

Engineered an AI-driven chatbot leveraging NLP to enhance customer support automation, handling real-time inquiries and product recommendations.

OpenAI GPT NLP React Python

Features

  • ๐Ÿค– Real-time Customer Support
  • ๐Ÿ’ฌ Natural Language Processing
  • ๐Ÿ›๏ธ Product Recommendations
  • ๐Ÿ“ฑ Multi-platform Integration
  • ๐Ÿ“Š Analytics Dashboard

Technologies Used

  • AI & NLP: OpenAI GPT, spaCy, NLTK
  • Backend: Python (Flask, FastAPI)
  • Frontend: React.js, Next.js
  • Cloud: AWS Lambda, Firebase
  • Database: MongoDB, Redis

Cloud Banking Platform

Developed a full-stack banking web application with a robust CI/CD pipeline and deployed on Azure Cloud, featuring secure authentication and transaction processing.

Azure React Python Docker

Features

  • ๐Ÿ”’ Secure Authentication
  • ๐Ÿ’ณ Account Management
  • ๐Ÿ’ธ Transaction Processing
  • ๐Ÿ“Š Analytics Dashboard
  • ๐Ÿ”„ CI/CD Pipeline

Technologies Used

  • Backend: Python (FastAPI), Node.js
  • Frontend: React.js, TailwindCSS
  • Cloud: Azure, Docker, Kubernetes
  • Security: OAuth2, JWT
  • CI/CD: GitHub Actions, Azure DevOps
In Development

LearnMate

An AI-powered study platform that personalizes learning content and provides intelligent recommendations. Features a ChatGPT learning assistant for interactive help.

TensorFlow FastAPI Streamlit ChatGPT

Features

  • ๐Ÿค– ChatGPT-Powered Learning Assistant
  • ๐ŸŽฏ Personalized Learning Recommendations
  • ๐Ÿ“Š Performance Analytics
  • ๐Ÿ”„ Cross-Model Validation
  • ๐ŸŒ Multi-Platform Integration

Technologies Used

  • ML & AI: TensorFlow, Scikit-Learn, Collaborative & Content-Based Filtering
  • NLP: ChatGPT, spaCy, NLTK, Hugging Face Transformers
  • Cross-Model: DeepSeek
  • Backend: Python (FastAPI, Flask)
  • Frontend: Streamlit
  • APIs: OpenAI (ChatGPT), Google Books, YouTube Data, Deepseek

Triplet SAE Project

Research pipeline comparing sparse autoencoder feature steering with LoRA fine-tuning for factual subject-predicate-object triplet extraction.

Python Gemma 3 SAE LoRA

Research Focus

  • Factual triplet extraction with qualifiers
  • SAE activation analysis and feature ranking
  • Ablation and steering experiments
  • Gemma 3 4B instruction-tuning with LoRA
  • Quantized model export for deployment

Technologies Used

  • Models: Gemma 3 4B IT, sparse autoencoders
  • Training: Hugging Face, PEFT, Accelerate
  • Evaluation: precision, recall, F1, entity overlap
  • Deployment: LoRA adapter, GPTQ 4-bit and 8-bit exports
GitHub Hugging Face

OLMo SAE Builder

Readable sparse autoencoder tooling for OLMo-3-7B-Think, built as a plain PyTorch single-GPU pipeline for activation collection and SAE training.

Python PyTorch OLMo SAE

Pipeline

  • Download OLMo-3-7B-Think locally
  • Stream text and collect residual activations
  • Compute pre-bias activation means
  • Train TopK sparse autoencoders from saved chunks
  • Single-GPU scripts designed for readability

Technologies Used

  • Model: allenai/Olmo-3-7B-Think
  • Training: PyTorch, bfloat16, forward hooks
  • SAE: tied decoder, TopK activation, reconstruction loss
  • Data: Dolma and Dolci-Instruct-SFT activation streams
GitHub Hugging Face
Paper Coming Soon

Brain-LLM Feature Alignment

Bachelor thesis testing whether brains and language models use overlapping transcript-grounded SAE features during naturalistic story comprehension.

fMRI SAE Ridge RSA

Thesis Question

  • Do brains and language models rely on overlapping transcript-grounded features?
  • Compare feature salience, predictive recoverability, and representational geometry
  • Focus on Narratives stories: shapesphysical and shapessocial
  • Finding: commonality appears stronger in which features matter than in shared geometry

Methodology

  • Models: Gemma 2 2B, Gemma 2 9B, Llama 3.1 8B
  • Brain data: OpenNeuro Narratives, Schaefer-200 cortical parcels
  • Pipeline: transcript SAE discovery, TR-grid alignment, ridge regression
  • Evaluation: held-out fit, feature-importance agreement, weight similarity, RSA

Skills & Expertise

Programming & Development

Languages

  • Python (Advanced)
  • C++ (Intermediate)
  • JavaScript (Intermediate)
  • SQL (Advanced)

Backend Development

  • Python (FastAPI, Flask)
  • Node.js (Intermediate)

Frontend Development

  • Vue.js

Cloud & DevOps

  • Azure
  • Bicep (Intermediate)

Infrastructure Tools

  • CMake
  • Makefile

CI/CD & DevOps

  • GitHub Actions
  • Azure DevOps
  • Docker & Containerization
  • Kubernetes Orchestration
  • Jenkins Pipeline Automation
  • Infrastructure as Code (IaC)
  • Automated Testing Integration
  • Continuous Deployment

Game Development

  • Object-Oriented Programming (OOP)
  • Game Mechanics Design
  • SFML for Graphics Rendering

Machine Learning & AI

  • Natural Language Processing (NLP)
  • Scikit-Learn, TensorFlow
  • AI-based Security Models
  • Recommendation Systems
  • CrewAI for AI Chatbot Development

Data Processing & Visualization

  • Power BI, Tableau
  • Pandas, NumPy
  • MySQL, PostgreSQL, MongoDB

Software Testing & Automation

  • ML Model Testing & Hyperparameter Tuning
  • Debugging & Data Pipeline Optimization
  • Web Scraping & API Integration

Computer Skills

  • Photoshop (Intermediate)
  • Social Media Management

Upon Request Websites

Custom websites developed for clients and businesses

SฮžNSฮž TRAVEL AGENCY

Travel Agency

Main page for a travel agency featuring modern design, social media integration, WhatsApp contact, and form submissions for inquiries.

React Next.js Tailwind CSS Vercel
๐Ÿ“ฑ Responsive Design ๐Ÿ’ฌ WhatsApp Integration ๐Ÿ“ง Contact Forms ๐Ÿ“ฑ Social Media

Reusa

Dominican Startup

Landing page for a Dominican startup showcasing their mission, services, and value proposition.

React Next.js Tailwind CSS Vercel
๐Ÿš€ Startup Landing ๐Ÿ’ผ Service Showcase ๐Ÿ“ˆ Growth Focused

Dr. Evangelina Soler

Medical Practice

Professional website for a medical doctor featuring patient information, services, and appointment booking.

HTML CSS JavaScript Responsive
๐Ÿ‘ฉโ€โš•๏ธ Medical Practice ๐Ÿ“… Appointment Booking ๐Ÿฅ Patient Information

Get In Touch

damorasoler@gmail.com
+34 614 21 17 93
Madrid, Spain