Startup News
Lambda (λ) Automata

Since late 2021, I've joined λ-automata as a Co-Founding Engineer.

We're building AI-first systems for autonomous surveillance, reconnaissance, and superhuman situational awareness.

We are a team of engineers, defense experts and former military officers, set on a mission to create cutting-edge technology for the European Defense sector at Silicon Valley-style speed and agility.

We're hiring.




Summary

Curiosity and results-driven engineer with a strong focus on applied machine learning research. Analytical thinker and methodical regarding good software, statistics and data engineering practices. Active Open source contributor and researcher. I stay up to date with ML research and continuously refine my skillset in order to have a good grasp of the industry's state and best practises. My skillset is focused on machine and deep learning, but it is also spans the dimensions of Web & Mobile App Development, AI on the Edge, IoT and Big Data.


Formal Education

  • Master of Computer Engineering and Informatics, University of Patras - (2016 - 2021)

  • Modules include: Machine Learning, Computer Vision, Data Engineering, Web Technologies, Software Engineering & Testing


Skills

CategoryTechnologies
LanguagesPython, C#, JavaScript, Java
DatabasesMySQL, MongoDB, TimescaleDB, SQLite
Machine Learning Keras & TensorFlow, PyTorch, Model Optimization, Distributed Training & TPUs, Apache Beam, Scikit-Learn, Jetson Platform, NVIDIA DeepStream
Software Engineering Design Patterns, Async Programming, ORM, CQRS, Testing, REST APIs, Distributed Systems
Web FrameworksASP.Net Core, Spring Boot, FastAPI, Vuejs, Grafana
DevOps and OtherGit, Gradle, CI/CD, Docker, Kubernetes

Github AccountMedium Blog


Work Experience

Research Engineer | Greek Army - Center of Technology and Research (Mandatory Army Service, 2021)


ML Research Engineer | Loctio.com (Contract, 2020 - 2021)

  • Collaborated with VLSI, Ground Position Systems and Software Engineers to evaluate and develop a satellite segmentation model (PointNet Variant) as part of a larger cloud based location estimation system. Increased past published research accuracy from 92% to 99.5% by applying proper training and preprocessing techniques

Responsibilities included:

  • Exploratory data analysis to determine if a machine learning model is suitable for the task at hand
  • Re-Implementation and improvement of existing papers
  • Prototyping with newer research advancements in order to improve model performance
  • Model optimization in order to deploy to embedded edge devices
  • Hardware-specific dataset creation

Researcher, Lecturer, Lab Assistant | University of Patras (2019-2020)

  • Performed research on: CNN Compression, Quantization, Semantic Image Inpainting with GANs
  • Lectured on Computer Vision and Deep Learning: Introduction to Deep Learning for Vision CNNs, Transformers and popular architectures Transfer Learning, Contrastive Pre-training GANs
  • Built lab exercises on basic computer vision tasks: classification, object detection, siamese network similarity, GANs

Android Developer | Eventora (Summer 2016)

  • Worked on the Ticket Scanner and Wallet apps

Machine Learning Projects

Cryptocurrency Trading Bot (2019 - Present)

  • Created an end to end ML Pipeline System to train a multivariate time series model based on market data and other features such as blockchain analytics and sentiment indicators
  • Crafted a data-acquisition/scraping service leveraging TimescaleDB
  • Implemented backtesting strategies with grafana dashboards and real-time slack notifications
  • Deployed everything mentioned in a Kubernetes cluster

PoseNet Dino (2020)

  • Used a pretrained pose estimation tensorflow lite model in order to play google chrome's "no internet" game by physically jumping and crouching

Highway Traffic (2020)

  • Used relative vehicle positions to train a recurrent neural network that predicts vehicle collision danger and congestion level
  • Leverage a pretrained vehicle detection model to detect shockwaves on highways
  • Train a small team of curators on dataset labelling

Shop & Item Sales Forecasting (2020)

  • Trained a recurrent neural network to forecast sales per shop, per item

Generative Test RNN Experiments on Discord Chat Logs (2019)

  • Trained a generative seq2seq LSTM model with teacher forcing to generate text from ~15 MB discord chat logs
  • Leveraged fasttext word embeddings for transfer learning
  • Trained a character-level GRU RNN to generate Greek song lyrics

Open Source

Fluent TFX (2020)

PyPi Package
  • Fluent TFX provides a better API for TFX pipeline creation. If you are already using tensorflow or keras for your models, this is an easy to use api that will have your model up and running in an e2e scenario quickly, without the need to waste a significant amount of time of learning the internals of tfx to get something working.

A Complete Autonomous Drones Study Plan (2019 - 2020)

  • I originally created this as a short to-do list of study topics for becoming a software engineer, I am creating this list to share my 3 year list of topics that I studied on the side during the curriculum of a Computer Engineering and Informatics degree. This is a list of short and medium length study topics to obtain knowledge regarding autonomous rotorcraft. The items listed here will give you enough knowledge to be able to understand how they work, their limitations and effort required to make them fly.

UAVCAN / libuavcan - Yukon (2019)

  • Architected a flask backend and a vue frontend solution, including the usage of server sent events
  • Optimised the transmission queue by simplifying frame transfer logic and most importanly, swap out a bottlenecking linked list queue with a custom AVL-Tree implementation, which supports equal keys, by using a custom block allocator to avoid fragmentation, in order to improve worse case performance by a large margin
  • Collaborated with the hardware and python api team to design the new web-based bus inspection tool

src-d / go-mysql-server (2019)

  • Implement some SQL functions to comply with the MySQL standard (unit tests included)
  • Make some parts of the query parser more robust


Other

Geolocation based Social Network App Frontend for Android (Java, Kotlin rewrite)

  • Implemented an offline-first app by using Room & Retrofit
  • Used MVVM with a mix of Observables and LiveData
  • Performed cross fragment/activity communication by EventBus
  • Extensively used Dagger 2, Architecture Components & the Design support library
  • Consume a Rest API & handle authentication with Jason Web Tokens
  • Features Include: Login/Signup, Post viewing, Likes, Comments, Friends & Blocking, Post Uploading (with image)

Game & Game Server Development (2011 - 2016 | Java, C++) - Freelancing

  • Published a game on Steam, made in Unreal Engine 4 with C++ (4000+ Sales)
  • Unreal Engine 4 contribution by implementing the Android local notification system
  • Created a series of game server modifications for a grid of Minecraft servers (400 concurrent players at peak, Java)


Hobbies

  • Express love for Software Engineering by contributing to Open Source projects (Hacktoberfest!) and Publishing of Software Engineering Articles on medium.com (Averaging ~6K article views per month)
  • Voxxed Days Athens 2018 Volunteer (3 consecutive years)
  • Photography, Audiovisual Arts through applied computer engineering