loraine.
b128564Blue Frontier2025 – nowSoftware Engineer6b76e40Arealec2022 – 2026Full Stack Developerfcbc1bfJack's Flight Club2022 – 2025Senior Full Stack Developer9061cc5University of Havana2017 – 2022Assistant Professor and Researcher8ea7b36Deepdatatech2021 – 2022Full Stack Developer

Loraine Monteagudo

Fullstack engineer with a machine learning background.

Now
Software Engineer at Blue Frontier
Based in
Tampa, Florida
Focus
Interfaces, APIs, data pipelines, infrastructure
Open to
Fullstack and platform roles

I like owning a product all the way down. I’ve built a simulator that models a year of an air conditioner’s energy use against real weather, a platform that dims a city’s street lights on schedule, and a pipeline that watches flight prices across six regions to find the ones worth telling people about. Each of those I built myself: the pipelines that clean and reshape the data, the API, the interface, and the model in the middle. Machine learning research is where I started, and it still does real work in what I build.

Email GitHub LinkedIn
6+ years of experience
30 projects built
52+ technologies worked with

Experience

2025.05 - present · 1 yr 3 mos

Software Engineer · Blue Frontier

Boca Raton, Florida · Remote · Full-time
  • Built two internal tools (SvelteKit + FastAPI on Azure) for a maker of ultra-efficient smart air conditioners: a sales simulator and a fleet operations console.
  • The sales tool runs a year of a building’s energy use against real past weather, compares the customer’s current AC with a Blue Frontier unit, and writes the proposal (PDF, Excel, Word).
  • Moved the core energy model from an old Excel/VBA workbook to Python (Pandas/NumPy): about 16x faster, and closer to the reference sheet.
  • The console watches and controls every unit in the field: live status, remote control, and commissioning runs that stream progress to the browser over WebSockets.
  • Added single sign-on with Microsoft Entra ID and role-based permissions to both apps, from login down to each endpoint.
  • Built ETL pipelines with Prefect over the fleet’s IoT sensor data in TimescaleDB: they clean raw readings, work out runtime, uptime, and health, and keep environments in sync.
  • Built a dashboard that measures and verifies energy savings: it pulls readings with SQL, cleans and groups them in Pandas, and fits regression models (scikit-learn) to each unit’s power use.
  • Kept Grafana as code for the whole fleet: dashboards, alerts, energy-saving counters, a map of every site, and export/import scripts to set it all up.
  • Wrote shared tools the company still uses: a Svelte component library on private npm, reusable GitHub Actions, and a library that checks the controllers inside each unit.
  • Ran infrastructure and delivery: Terraform on Azure, CI/CD pipelines that redeploy only what changed, secrets in Azure Key Vault, and OpenTelemetry across services.
  • Set up tests in four repositories, backend and frontend, running on every change.
PythonTypeScriptHTML & CSSLinuxDockerGitFastAPISQLAlchemyPostgreSQLTimescaleDBRedisSvelteTailwindCSSWebSocketsAzureTerraformPrefectGrafanaOpenTelemetryGitHub ActionsPandasNumPyscikit-learnPytest Ownership & initiativeCross-team collaboration
2022.02 - 2026.03 · 4 yrs 1 mo

Full Stack Developer · Arealec

Zaragoza, Spain · Remote · Contract
  • Built a smart street-lighting platform on LoRaWAN, from the lamp controllers in the field to the operator dashboard: lamps report energy use and status, and operators control one lamp or a whole group.
  • Wrote the backend in Python (FastAPI, SQLAlchemy, MySQL): nested installation trees, schedules and calendars with exceptions, work orders, alarms, energy reports, and a full audit log.
  • Wrote the device layer: a binary protocol small enough for low-power radio messages, plus multi-step device jobs that retry and recover.
  • Added over-the-air firmware updates: the system splits each image, sends it to a whole group of lamps at once, tracks progress, and recovers from failures.
  • Built a scheduler that cuts energy use: lamps switch at the true sunrise and sunset for their GPS location and dim through the night on set programs. If a lamp does not confirm, the system raises an alarm.
  • Built the Vue 3 dashboard: a map of every device, live status over WebSockets, manual control of lamps and groups, schedule and calendar editors, user management, alarms, and work orders.
  • Designed permissions that follow the installation tree: a grant on a group flows down to everything under it, so a city gives each contractor just what it needs.
  • Embedded Grafana charts in the app: energy use over time, lamp and gateway status, live alarms, and email alerts when something looks wrong.
  • Ran self-hosted error tracking (GlitchTip, Sentry-compatible) on every service. A new server sets up its own projects and alerts on first boot.
  • Ran the platform as Docker services behind an HTTPS reverse proxy, with GitHub Actions deploys to several environments, database migrations, backup and restore scripts, and written guides.
  • Built tests for protocol encoding, permissions, schedules, and work-order states. They run on every change.
  • Worked with the hardware engineers on the device protocol and made the link between cloud and field more reliable.
PythonSQLAlchemyJavaScriptHTML & CSSLinuxDockerWebSocketsGitFastAPIMySQLVueTailwindCSSMQTTgRPCRedisGrafanaGitHub ActionsSentryCaddyCloudflare Ownership & initiativeCross-team collaborationTechnical documentation

Jack's Flight Club · 2 yrs 4 mos

2024.09 - 2025.04 · 7 mos

Senior Full Stack Developer

Tampa, Florida · Remote · Full-time
  • Built a compiler for the shorthand deal writers use for routes, prices, and conditions: a grammar turns it into a typed tree that gives both the stored data and the text readers see, in place of a manual step that often went wrong.
  • Led the redesign of the platform’s core map data (airports, routes, regions). One fast central store replaced copies in two databases, which simplified the code and ended a whole set of data bugs.
  • Sped up deal scanning by tuning queries, batching writes, caching, and running work in parallel, so more data did not slow it down.
  • Set up Sentry on every service and tied it to deploys, so we found the cause of a fault much faster.
  • Automated database chores people used to do by hand: backups, cleanup of old data, and a weekly refresh of the test environment from production.
  • Automated Google Sheets reports and regional deal alerts, which cut hand work and caught high-value deals in time.
  • Looked after the machine learning model in production, fixing bad predictions and guarding against false positives.
  • Added new data sources, including sites behind anti-bot protections.
  • Worked closely with the deals team so the tools fit the way they work each day.
PythonTypeScriptDockerFastAPISQLAlchemySentryGoogle SheetsMLFlow Ownership & initiativeCross-team collaboration
2022.12 - 2024.09 · 1 yr 9 mos

Full Stack Developer

Tampa, Florida · Remote · Full-time
  • Built a flight-deal tracker as a set of small services: scrapers collect prices from many sources, a cleaning stage makes them consistent, and a detector compares each price with its past average to find real deals.
  • Built the data pipeline: it collects daily across 6 regions, cleans messy source data (airports, currencies, languages), and keeps every version.
  • Trained a machine learning model that judges whether a flight is a good deal and improved it with the team’s own calls as feedback. It took over much of the work people did by eye.
  • Built the data sets and training flow behind it: automatic extraction, human feedback added to the training data, and experiment tracking so we could repeat any model version.
  • Built an internal Svelte dashboard where the team sorts, checks, and publishes deals as they arrive, with a live view of what the system finds.
  • Built a Slack bot that became the team’s main tool: commands, forms, and daily updates that cut status checks by about half.
  • Automated delivery with GitHub Actions: tests on every change, one-click deploys, and over 20 scheduled jobs that run tracking, reports, and upkeep on their own.
  • Packaged every service with Docker so it runs the same in development and production.
  • Linked the platform to the rest of the business: the email marketing system, Google Sheets and BigQuery for reports, and AI help with editing deal write-ups.
  • Set up tests from the start, which made it safe to change and ship code fast.
PythonTypeScriptHTML & CSSLinuxDockerFastAPISQLAlchemyPandasPlaywrightSvelteRedisPostgreSQLPytestCloudflareGoogle BigQueryDockerGitHub Actions Ownership & initiativeCross-team collaboration

University of Havana · 4 yrs 11 mos

2022.01 - 2022.08 · 7 mos

Assistant Professor and Researcher

Havana, Cuba · On-site · Full-time
  • Taught Logic, Machine Learning, and Programming, and built the courses around hands-on projects and assignments that tie theory to real problems.
  • Worked on AutoGOAL, an open-source AutoML system that searches for machine learning solutions to any problem. Added meta-learning from past runs, so it starts closer to a good answer on new data, which cuts training time and raises accuracy.
  • Guided students through their AI research projects, from shaping the question to publishing the results.
PythonScikit-LearnPandasKerasGitC#Jupyter Notebooks Mentoring
2017.09 - 2021.12 · 4 yrs 3 mos

Undergraduate Student Researcher

Havana, Cuba · On-site · Part-time
  • Researched natural language processing: pulled structured facts out of medical text and helped build the labelled data sets behind the international IberLEF eHealth challenges.
  • Competed in the IberLEF eHealth Knowledge Discovery Challenge 2021 with strong results against international teams.
  • Presented research at the XVI COMPUMAT 2019 International Congress and won awards at school and national Student Congresses across several years.
PythonScikit-LearnPandasGitspaCyNLPNumPyTensorFlowKerasJupyter Notebooks
2018.09 - 2021.12 · 3 yrs 3 mos

Undergraduate Teaching Assistant

Havana, Cuba · On-site · Part-time
  • Helped teach Logic over 3 semesters, leading practice sessions and guiding students through hard concepts.
  • Helped set and grade exams, keeping marks fair and feedback quick and useful.
C# Mentoring
2021.06 - 2022.01 · 7 mos

Full Stack Developer · Deepdatatech

Havana, Cuba · Remote · Contract
  • Built a web app for psychological testing in hiring: candidates take personality tests based on Carl Jung’s theory, and the results guide the hiring decision.
  • Wrote the backend in Python (Flask) with MongoDB: it serves each test, scores it, and stores the results.
  • Built the candidate’s side with server-rendered templates (Jinja) and Vue where the page had to react, which kept it fast without a heavy framework.
  • Built a tool (Streamlit) so the team could write and publish new tests — questions, structure, and scoring — without code. Tried a CMS (Strapi) first and chose the lighter tool.
  • Worked remotely as a contractor, taking features from design to deployment.
PythonTailwindCSSJavaScriptFlaskHTML & CSSGitDockerLinuxStreamlitStrapi Ownership & initiative

Featured projects

All 30 →

Tools I work with

All skills →

LANGUAGES

Python TypeScript JavaScript C# Go HTML & CSS

BACKEND & DATABASES

FastAPI Flask ASP.NET Core SQLAlchemy PostgreSQL TimescaleDB MySQL Redis

FRONTEND

Svelte Vue React TailwindCSS

DATA & MACHINE LEARNING

Pandas NumPy Scikit-Learn TensorFlow Keras spaCy MLFlow NLP AutoML Streamlit Google BigQuery

DEVOPS & CLOUD

Docker GitHub Actions GitOps Terraform Azure Linux Git Aspire Portainer Caddy Prometheus Grafana OpenTelemetry Cloudflare Home Assistant

PROTOCOLS & MESSAGING

REST APIs WebSockets MQTT gRPC MCP

TESTING & MONITORING

Pytest Playwright Sentry

Education & achievements

BSc. Computer Science

University of Havana

2016-2021 · GPA 4.81 / 5

Member of the Artificial Intelligence research group, with publications, conferences and competitions.

CERTIFICATIONS

PUBLICATIONS

ACHIEVEMENTS

  • Silver Award National Computer Conference, annotation model for health-tech knowledge discovery
  • Gold Diploma 2018 Student Scientific Conference, ML negation detection in Spanish
  • Silver Diploma 2019 Student Scientific Conference, general-purpose annotation model
  • Conference publication XVI COMPUMAT 2019 International Congress, negation detection paper
  • 5th Place IberLEF eHealth Knowledge Discovery Challenge 2021
  • Scientific Merit Award For excellent academic and research results (undergraduate)
  • Gold Diploma Graduated with top academic scores