Alejandro Miranda Franco.

Solutions architecture and applied artificial intelligence

Solutions Architect

ML/AI Engineer · Full-Stack Developer · Cloud Solutions Architect

With 6 years of experience in software development (4 focused on ML/AI), I build full-stack products and AI solutions. My work combines cloud architecture and multimodal processing through deep learning, reinforcement learning, NLP, and computer vision (including OCR). In generative AI, I adapt LLMs and orchestrate multi-agent systems using RAG and semantic search approaches. I contribute throughout the full software lifecycle—from requirements and architecture to deployment and operations—working closely with multidisciplinary teams and executives. I completed my studies in Applied Mathematics and Computing and completed diploma programs in Data Science, AI Engineering, and Deep Reinforcement Learning; I am currently studying Applied LLM Engineering and Agentic AI.

I enjoy solving complex problems.

Technologies I work with

Python
TensorFlow
PyTorch
scikit-learn
JavaScript
React
Vue.js
Astro
Laravel
Go
C
C++
Node.js
Django
FastAPI
MySQL
PostgreSQL
MongoDB
Supabase
AWS
AWS Amplify
Azure
GCP
Git
Terraform

Technical skills

Full-stack development & DevOps/MLOps

Full-stack: JavaScript, PHP, Java, Go, C/C++ and SQL/NoSQL. React, Next.js, Vue.js, Nuxt, Astro, Node.js, Django, FastAPI, Laravel and Apache Kafka.
Methodology: Waterfall and Scrum.
APIs & messaging: REST, GraphQL, webhooks, gRPC, Server-Sent Events and Pub/Sub.
DevOps/MLOps: CI/CD, Continuous Testing (CT), Git, GitHub and Azure DevOps.

Machine Learning & AI

Libraries: Python, TensorFlow, TensorFlow.js, Keras, PyTorch, scikit-learn and Gymnasium.
Models and applications: Convolutional neural network (CNN), optical character recognition (OCR), speech-to-text, text-to-speech and natural language processing (NLP).
Deep Reinforcement Learning: Bandits, Monte Carlo, Markov decision processes (MDP), dynamic programming, exploration and exploitation, agents, environments and policy gradients. DQN, Double DQN (DDQN) and Dueling DQN. SARSA, n-step SARSA, Q-learning, Dyna Q, eligibility traces and TD (λ).
AI development: Architecture, data creation and engineering, model training, evaluation and inference, and deployment.

Generative AI & data

Data: Medallion architecture, embeddings and vector search with FAISS.
Generative AI: Gemini API, ChatGPT API, CLIP, diffusion models, large language models (LLMs), intelligent agents and multi-agent systems, agentic AI, retrieval-augmented generation (RAG) and semantic search.
Data Warehouse: Snowflake, Databricks and BigQuery.
Vector databases: Pinecone and Pinecone CLI.

Cloud & infrastructure

Cloud: AWS, Azure, GCP, Vultr and Contabo. Infrastructure as code with Terraform.
CLI: AWS CLI, gcloud CLI and Azure CLI.
BaaS & tools: Amplify (Gen 1 and Gen 2), Firebase, SWA CLI and Supabase.

AI engineering & automation

AI-assisted software engineering: Spec-Driven Development (SDD), harness engineering for coding agents, multi-agent orchestration, quality gates and verification workflows.
Automation: n8n, Make and custom integrations with Telegram, WhatsApp, Gmail, Google Calendar, Drive, Facebook, Instagram, LinkedIn, TikTok, YouTube and other platforms.

Security

Security: Experience as an ISO 27001 internal auditor.
Cryptography: Symmetric and asymmetric encryption, hash functions and encoders.

01 PYTHON 1001 RAG
{ CLOUD : ML / AI }
0101 FULL-STACK 0110
SDD 0101 TENSORFLOW
{ DATA : SOLUTIONS }
1010 CODE 0110 1001

Experience.

Nov 2025 — Present
Intelligent Solutions Architect
SM Ingeniería
I design solutions to automate internal administrative, operational and logistics processes, from requirements definition and cloud architecture through implementation and deployment, in collaboration with technical, operational and executive teams.
Jul 2022 — Aug 2025
Software Developer / ISO 27001 Internal Auditor
Certificación Digital Online
I designed and deployed a custom CNN. I developed liveness verification, ID-based electronic signing and electronic-signature workflows in collaboration with technical, operational and executive teams, and participated as an internal auditor in the 2024 ISO 27001 recertification.
Jan — Jul 2022
Software Developer
Interstack
I built internal tools to monitor digital campaigns and conducted security assessments to support executive decisions.
2020 — 2022
Full-Stack Developer
FES Acatlán, UNAM
I built a web system to manage drinking-water tests, with a view for users, laboratory result uploads and reports for university leadership.
2020 — 2022
Full-Stack Developer / Cloud Solutions
SM Ingeniería
I built the corporate website, internal tools and cloud deployments; I automated Facebook, Instagram and LinkedIn publishing with Make.

Education.

2026
Applied LLM Engineering
Actumlogos
2026
Agentic AI
Actumlogos
2026
Deep Reinforcement Learning
Actumlogos
2026
AI Engineering Diploma Program
Augmented Humans
2022
Data Science Diploma Program
FES Acatlán, UNAM
2018 – 2021
Applied Mathematics and Computing Degree Studies
FES Acatlán, UNAM

Applied LLM Engineering · Learn moreShow less

Natural language processing (NLP)
Information retrieval
Introduction to Applied LLM Engineering
Clinical data preprocessing and candidate filtering
Clinical matching search
Scoring system
Introduction to trading bots
Financial analysis
Sentiment analysis in trading
Deep learning for stock markets
LLMs for stock market bots
Reinforcement learning for stock markets
Trading bot deployment

Agentic AI · Learn moreShow less

AI, agents and environments fundamentals
Agent architectures, rationality and utility functions
Chain of Thought, ReAct, Self-Refine and Self-Consistency
Knowledge representation: logic, rules and ontologies
Bayesian networks, uncertainty and problem solving
Informed and local search: A*, heuristics and genetic algorithms
Supervised, unsupervised and reinforcement learning
Neural networks and deep learning: CNNs, RNNs and Transformers
LLMs, Transformer architecture and prompt engineering
Fine-tuning, LoRA and Instruct Tuning
RAG and search across your own documents
Self-Refine, Test-Time Compute and hallucination control
Multimodal image and audio agents
Multi-agent systems with LangChain, CrewAI and Ollama
Local, cloud and API deployment
Python, PyTorch, TensorFlow, Keras, Jupyter, Hugging Face and FastAPI
Ethics, privacy and bias

Deep Reinforcement Learning · Learn moreShow less

MDP
POMDP
Monte Carlo
Dynamic programming
TD methods
SARSA
Q-learning
Dyna-Q
DQN
Double DQN
Dueling DQN
REINFORCE
Actor-Critic
A2C
TRPO
PPO
DDPG
HER
SAC
PyTorch
Gymnasium

AI Engineering Diploma Program · Learn moreShow less

Multimodal data engineering
Medallion architecture
Pub/Sub
Embeddings
CLIP
PCA
t-SNE
Latent spaces
Deep learning
CNN
Transfer learning
Transformer architecture
LLM
RAG
Grounding
Agentic AI
Google ADK
MCP
A2A
DevOps
MLOps
Production deployment

Data Science Diploma Program · Learn moreShow less

Data quality
Data ingestion
Supervised classification
Supervised regression
Dimensionality reduction
Clustering
Neural networks
Convolutional neural networks

Applied Mathematics and Computing Degree Studies · Learn moreShow less

Machine learning
Data mining
Forecasting
Distributed databases
Probability
Univariate and multivariate statistics
Stochastic processes
Stochastic simulation
Linear algebra
Optimization
Univariate and multivariate calculus
Differential equations
Numerical methods
Algorithm analysis
Discrete mathematics
Cryptography
Modular algebra
Number theory

Press.

Recognition · MIT COVID-19 Challenge

Track A winner — Latin America vs. COVID-19

In 2020, I was part of the team that developed a solution to assess COVID-19 immunity and reinfection risk. I contributed by designing and developing the product’s web architecture.

A collective result developed with members from several Mexican institutions.

Year: 2020
Winning track: Track A
Participants: 1,500
Countries: 73

Interviews, video, and reporting on the recognition.

Contact.