Hi. I'm

Lucas
Cardoso.

Senior Software Engineer · AI & Computer Vision

I build AI systems that work outside the notebook: on factory machines, in irrigation canals, at the edge.

Today I develop embedded computer vision at Delta Máquinas Têxteis. Before that, I founded and led Green Next for over six years.

Taking on custom projects. Computer vision, IoT, Python automation and R&D project management, alongside my current role. Remote, fixed scope, limited capacity.

Limited availability. I have one project underway and will open a new slot soon. Send your idea now to book the conversation.

Fully booked right now. I'm not taking new projects. Leave your contact and I'll let you know when a slot opens.

Photo of Lucas Cardoso, senior software engineer in AI and computer vision
Lucas Cardoso · Blumenau, Brazil
50+IoT sensors in the field, on commercial rice farms
~30%energy savings in pumping
15people on the team at Green Next's peak
R$ 1.2M+raised in investment and R&D funding

Green Next figures, 2019–2025.

About me

Engineering that leaves the lab.

Automation engineer by training, M.Sc. candidate in Computer Engineering, with six years leading an agritech startup.

I work on object detection with YOLO, image processing and machine learning pipelines that run in production, using PyTorch, TensorFlow and OpenCV. I like the hard part: getting the model onto the hardware, into the field and onto the production line.

As founder and CEO of Green Next, I led the development of Hydra, an IoT platform for irrigation management, from 2019 to 2025. In my M.Sc. at FURG, I research irrigation forecasting and simulation with deep learning, using real data from those devices.

I specialize in R&D project management: I track technology maturity on the TRL scale and run execution with agile methodologies, so every stage ends in something testable, not just a report.

Custom projects

Need AI that works in the real world?

I take on outside projects alongside my current role. Only a few at a time, so I can deliver them with the same care I put into my main job.

  • Remote, with meetings outside business hours.
  • Fixed scope and price per stage, with weekly deliveries.
  • You own the code and documentation at the end of the project.
  • Contracted and invoiced through Zenith Softworks, my company.
  • I don't take projects in the textile industry, to avoid any conflict with my employer.
Tell me your idea Join the waitlist

Computer vision

Detection, counting, inspection and image reading with YOLO and OpenCV, from dataset to a model running at the edge.

Track record: embedded visual inspection on industrial machinery

IoT and telemetry

LoRa sensor networks, field data collection and storage, and remote equipment actuation.

Track record: Hydra, 50+ sensors on commercial farms

Predictive models

Time series forecasting and process simulation from sensor data, with honest evaluation.

Track record: M.Sc. research on irrigation (FURG)

Python automation and dashboards

APIs in FastAPI or Django, dashboards in Dash or Streamlit, and routines that take spreadsheets and manual work out of the way.

Track record: Hydra's web platform and internal tools

R&D project management

Planning and control of innovation projects: milestones on the TRL scale, execution in agile cycles and reports that show maturity progress, from concept to validated prototype.

Track record: R$ 719k in R&D with SEBRAE, EMBRAPII and ITEC-FURG

TRL scale
  1. 30-minute call

    Free. I learn about the problem, the data you already have and whether it makes sense to move forward.

  2. Fixed proposal

    Within a week: scope, stages with TRL milestones, timeline and price per stage.

  3. Weekly deliveries

    Something working to look at every week, not just at the end.

  4. Delivery and handover

    Code, documentation and a session for your team to take over.

Case studies

Real projects, real results.

What was built, how, and what changed.

Sensors level · flow Pumps actuation Gateway LoRa Cloud data · rules App & dashboard grower
Green Next · 2019–2025

Hydra

Problem
Irrigated rice farms waste water and energy running pumps on fixed schedules, without knowing the actual water level in the canals.
Architecture
Sensors in the irrigation canals and pump telemetry, LoRa communication, a cloud platform, an offline-first mobile app and real-time reports for the grower.
Result
From idea to production on commercial rice farms, with large producers and cooperatives: 50+ sensors in the field, around 30% energy savings and a team of up to 15 people.
  • IoT
  • LoRa
  • Python
  • Django REST
  • PostgreSQL
  • Agribusiness
Dashboard Controller Pump command status & consumption
R&D · SEBRAE, EMBRAPII and ITEC-FURG

Remote pump actuation

Problem
Switching irrigation pumps on and off means driving out to the field, often in the middle of the night and kilometers away.
Approach
Remote actuation and monitoring technology for the pumps, integrated with Hydra, with water and energy consumption forecasts from sensor and weather data.
Result
A R$ 719k R&D project with SEBRAE, in partnership with FURG's EMBRAPII unit and ITEC-FURG.
  • Automation
  • PLC
  • IoT
  • R&D management
measured forecast Illustrative
M.Sc. · FURG · 2025–2027

Irrigation forecasting with deep learning

Question
Can we forecast and simulate irrigation dynamics in rice cultivars from the sensors already installed in the field?
Approach
Deep learning models for time series, such as N-BEATS, trained on data from Hydra devices, IoT sensors and weather data. Advisor: Prof. Dr. Adriano Velasque Werhli.
Goal
Use less water and energy on the farm by deciding when to irrigate based on what will happen, not on the clock.
  • Python
  • Deep learning
  • N-BEATS
  • Time series
In progress on GitHub
iot-ops-agentcoming soon

Assistant that answers questions about sensor telemetry and technical manuals, with RAG and human approval.

cv-mlops-pipelinecoming soon

Object detection from training to deployment: MLflow, ONNX, a FastAPI service and latency metrics.

Tech stack

By area.

A new item only goes on this list once there is a public project to back it up.

Computer vision

  • YOLO
  • OpenCV
  • PyTorch
  • TensorFlow
  • ONNX · in progress

Data and ML

  • Python
  • R
  • Pandas · NumPy
  • Scikit-learn
  • MLflow · in progress

IoT and automation

  • LoRa
  • Field sensors
  • PLC
  • C

Backend and deployment

  • FastAPI · Django REST
  • PostgreSQL
  • Docker
  • AWS
  • Plotly Dash · Streamlit

R&D management

  • TRL scale
  • Agile methodologies
  • Projects with SEBRAE and EMBRAPII
Career

Experience.

  1. Nov 2025 – Present

    Senior Software Engineer, AI & Computer Vision

    Delta Máquinas Têxteis · Pomerode, SC

    Developing the AI-based visual inspection system for Delta's textile machinery: computer vision models embedded in the machine that identify fabric defects in real time. Technical details are confidential.

  2. 2019 – 2025

    Founder & CEO

    Green Next · Rio Grande, RS

    Founded and led an agritech startup that automates water management for irrigated crops. Took the Hydra platform from idea to production on commercial rice farms: 50+ IoT sensors, around 30% energy savings and a team of up to 15 people. Raised R$ 500k in investment (Ventiur and a private fund linked to Agropecuária Canoa Mirim) and R$ 719k in R&D funding with SEBRAE, EMBRAPII and ITEC-FURG.

Studies

Education.

  1. 2025 – Feb 2027 (expected)

    M.Sc. in Computer Engineering

    Universidade Federal do Rio Grande (FURG)

    Provisional thesis title: "Modeling and Simulation of Irrigation in Rice Cultivars". Deep learning forecasting (N-BEATS) from Hydra field data. Advisor: Prof. Dr. Adriano Velasque Werhli.

  2. In progress · until May 2027

    Postgraduate Specialization, Software Engineering for Applied AI

    UNIPDS

    LLMs, RAG, agents and software engineering with AI in production.

  3. 2013 – 2020

    B.Eng. in Control and Automation Engineering

    Universidade Federal do Rio Grande (FURG)

    Final project: experimental analysis of LoRa technology for remote sensing on rural properties, with point-to-point tests from 240 m to 12 km. Advisor: Prof. Dr. Vitor Irigon Gervini.

Languages: Portuguese (native), English (full professional), Spanish (limited working).

Available for new projects Limited availability · next slot opening soon Fully booked · waitlist open

Let's work
together.

A computer vision, IoT, automation or R&D project? A 30-minute call is enough to see if it's a fit.

To help me reply faster, tell me:

  1. The problem and who is affected by it today
  2. What data or images already exist
  3. Timeline and budget range, if you have one
lucas@znsoftworks.com