Lucas Scheinkerman

Robotics, Computer Vision & AI Engineer

Robotics · Edge AI · LLMs · Python & C++

I build robotics, computer vision and LLM-based systems, from model integration and fine-tuning to edge deployment, monitoring, and production applications.

LinkedIn Github Upwork Contact Me Get my CV


Buenos Aires, Argentina
University of Buenos Aires

About Me

Lucas Scheinkerman

I’m a Robotics, Computer Vision & AI Engineer who builds systems that move from prototypes into real-world environments: NVIDIA Jetson devices, industrial equipment, and robotics simulations.


I work end to end across model integration and task-specific fine-tuning, application development, deployment, and monitoring. Currently, I’m a Founding AI & Robotics Engineer at IAMAI Simulations, building simulation infrastructure to train, test, and validate drone autonomy systems.


My work includes embedded AI assistants, edge computer-vision pipelines, and simulation infrastructure for autonomous systems. I enjoy turning complex research and platform constraints into dependable engineering systems.


Based in Buenos Aires, I work in English, Spanish, and Hebrew.

Tech Stack

Languages

C++
Python
Java

Computer Vision

PyTorch
OpenCV logo OpenCV
CUDA logo CUDA
Ultralytics YOLO logo YOLO · DETR · Vision Transformers

AI Systems

LLMs & Fine-tuning
RAG
ASR & TTS

Robotics & Edge AI

Unreal Engine
ROS
Android
Docker
Google Cloud
BalenaOS
Git
CMake
Bazel
NVIDIA logo NVIDIA Jetson
Grafana logo Grafana
Prometheus logo Prometheus

Languages

Spanish Spanish
English English
Hebrew Hebrew

Work experience

Founding AI & Robotics Engineer

IAMAI Simulations
May 2026 - Present

As a founding engineer of a very small company, I work wherever it's needed most, from implementing core features and improving developer workflows to supporting our open-source community and helping shape technical direction.

Responsibilities:

  • Lead the implementation of core features across our drone simulation platform.
  • Maintain our open-source repositories and review community contributions.
  • Mentor and supervise junior engineers across multiple projects.
  • Integrate AI models and AI-driven workflows into the platform to support autonomous behaviors, perception, and developer tooling.
  • Designed and maintain our CI/CD pipeline, reducing review overhead and deployment friction.
  • Provide technical support to the community through GitHub and Discord.
  • Prepare and deliver technical demos for potential clients.
  • Collaborate on technical planning, architecture, and product decisions.
Project AirSim source Simulation documentation

AI & Robotics Engineer

Marvik
January 2026 - May 2026

Designed and implemented an embedded virtual assistant for industrial heavy machinery, including loaders, excavators, and aerial lifts.

  • Combined ASR, RAG, and TTS into a speech-enabled assistant experience.
  • Fine-tuned language models for domain-specific applications.
Demo video CES 2026 post

Computer Vision Engineer

Optriment
January 2025 - January 2026
  • Developed computer-vision pipelines with PyTorch, OpenCV, CUDA, Python, and C++.
  • Deployed AI services to Jetson Nano edge devices and managed the C++ build infrastructure.
  • Built monitoring dashboards with Grafana and Prometheus; deployed services with Docker Compose and GCP.
  • Fine-tuned vision-transformer models for task-specific applications.
Computer Vision in a Box

Robotics simulation Engineer

Unreal Empowerment
June 2023 - January 2025

Developed simulation tools to generate artificial datasets for training AI-based aircraft navigation algorithms, ensuring compliance with specific requirements.

Used C++ for core development and Unreal Engine for high-fidelity visual rendering, enhancing the realism of simulations.
Also I employed Python for implementing ROS nodes and handling API client calls, effectively integrating various technologies to produce robust and accurate training environments.

Infrastructure Software Engineer

Ekumen
September 2019 - May 2022
  • Designed and implemented data collection and ingestion pipelines using Python’s Apache Beam SDK, optimizing data processing workflows.
  • Developed Android applications focused on sensor data collection, augmented reality (AR) features, and live data plotting using Android Java SDK and C++.
  • In charge of code integration, testing, maintenance, and review to ensure high-quality software.
  • Mentored new hires in introductory-level C++ training, enhancing their foundational programming skills.
  • Conducted technical interviews, assessing candidate qualifications and technical expertise.
  • Participated in research group using ROS and SLAM techniques.

Project AirSim-Based Solutions

Simulation, integration, and validation workflows for autonomous systems built on the open-source Project AirSim platform.

Project AirSim ground vehicles in a simulated environment Project AirSim X8 fixed-wing aircraft with simulated sensor outputs

From autonomy prototypes to repeatable simulation workflows

I work on simulation infrastructure for drone autonomy systems, helping teams configure environments, vehicles, sensors, controllers, and client integrations for development and validation.

  • Configurable drone, fixed-wing, and ground-vehicle simulation scenes.
  • Unreal Engine environments with rendered sensors and high-fidelity world geometry.
  • Python, C++, and ROS 2 integrations for autonomy-stack development.

Selected Case Studies

Selected technical work across industrial AI, edge computer vision, and applied model fine-tuning.

Technicians using tablets on an industrial shop floor

Embedded Industrial AI Assistant

Designed an assistant for heavy equipment environments, combining speech recognition, retrieval-augmented generation, text-to-speech, and domain-specific language-model fine-tuning.

ASRRAGTTSLLM Fine-tuning
Golf swing analysis application interface

Golf Swing Analysis App

An advanced golf swing analysis system using MediaPipe pose detection, OpenCV, and NumPy. It identifies swing phases with a custom-trained vision-transformer classifier and provides annotated visualizations.

Built with an extensible architecture, REST-ready FastAPI interface, CLI, structured logging, and pytest coverage. The deployment uses AWS Lambda with RunPod GPU, and an LLM chat layer provides AI-powered swing feedback through OpenAI, Gemini, and Grok.

MediaPipeOpenCVVision TransformersFastAPIAWS ServerlessLLM Chat

Padel Video Analytics | Player Tracking & Court Heatmaps

Built and deployed a Python analytics pipeline on NVIDIA Jetson that turns fixed-camera match footage into annotated video, player tracking data, and court heatmaps.

Integrated court calibration, 2D ball tracking, pose overlays, and player identity linking. Generated time-weighted heatmaps for all four players, including observation coverage and explicit handling of missing data.

PythonNVIDIA JetsonPlayer TrackingPose EstimationHeatmaps

Fabric Rolls Warehouse Analytics

An end-to-end computer-vision pipeline for indoor logistics and retail-style scenes, using fabric rolls as the primary inventory object and people as co-occurring actors.

Transforms ordered frame streams from disk, extracted video, or upstream capture into versioned, machine-consumable annotations and optional visual deliverables.

Computer VisionObject DetectionTrackingVideo Analytics