Skip to content
View esshka's full-sized avatar
🏠
Working from home
🏠
Working from home

Block or report esshka

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
esshka/README.md

Hi, I'm Eugene — Software Engineer & Architect

Headline: Software Engineer | Ruby/Python/Java/Clojure, Generative AI, Algo Trading, Event Sourcing/CQRS/DDD, Computer Vision, Real-time Systems, Modern UI.

Location: Serbia, Belgrade
Email: esshka@gmail.com
Web: esshka.github.io
GitHub: github.com/esshka
Twitter: twitter.com/genesshk

Professional Summary

Software Engineer & Architect with 15+ years of experience building complex, high-performance systems. Background in Applied Mathematics and Computer Science provides strong fundamentals in algorithms, data structures, and the theoretical foundations of machine learning.

Currently focused on Generative AI, Computer Vision, Algorithmic Trading, and Neuroevolution, leveraging deep expertise in Java and Clojure ecosystems for high-concurrency systems, alongside low-latency engineering and Python for Machine Learning models. Extensive track record in Fintech/DeFi, Real-time Metaverse, Distributed Microservices and Full-Stack development. A proponent of Clean Architecture, DDD, and Event-Driven systems, combining the hands-on delivery of a Senior Engineer with the strategic oversight of an Architect. Maintained Top Rated Plus status on Upwork since 2016 with a 100% Job Success Score.

Core Expertise & Technologies

Architecture & Patterns

Distributed Systems, Event-Driven Architecture (EDA), CQRS & Event Sourcing, Domain-Driven Design (DDD), Hexagonal/Clean Architecture, SOLID, Microservices, High-Frequency Trading (HFT) patterns, CleanML.

Frontend & UI Architecture

React Svelte Next.js MobX Redux Canvas WebGL

React, Svelte, Next.js, Advanced State Management (MobX/Redux/Zustand/Signals), Unidirectional Data Flow, Micro-frontends, Server-Side Rendering (SSR), Real-time Visualization (Canvas/WebGL).

Generative AI & Machine Learning

FastAPI LangChain LlamaIndex PyTorch TensorFlow Computer%20Vision RAG

Architecting Autonomous Agentic Systems and Pipelines, MCPs, Advanced RAG (Hybrid Vector/Graph Search), LLM Fine-tuning (p-tune, Lora, SFT), Neuroevolution, Reinforcement Learning (PPO, Q-Learning), Genetic Algorithms (Neuroevolution), Computer Vision, PyTorch/TensorFlow, Architecting custom models.

Languages & Runtimes

Python Ruby Java Clojure TypeScript JavaScript C%23 Go OCaml

Primary: Python (FastAPI, LangChain/LlamaIndex, Machine Learning), Ruby (Ruby on Rails - Full Stack dev), Java (High-Performance, Low-Latency GC, Concurrency), Clojure (Concurrent Systems, Stream Processing, Immutable Data)

Full-Stack/Web: Ruby (Ruby on Rails), TypeScript (React, Svelte), JavaScript (Node.js), Historic Mastery: C# / .NET Ecosystem (MCSD Certified)

Utility & Hobby: Go (System tools, Microservices), OCaml (Functional patterns, Prototyping)

Data & Storage

PostgreSQL SQLite Elasticsearch Kafka Redis ClickHouse pgvector Pinecone Milvus

PostgreSQL, SQLite, Elasticsearch/OpenSearch, Kafka, Redis, ClickHouse, Time-series DBs, Vector DBs (pgvector/Pinecone/Milvus).

Cloud & DevOps

Docker GitHub%20Actions Prometheus Grafana MLOps

MLOps (RAG, Inference pipelines), Docker, GitHub Actions, Observability (Prometheus/Grafana/Tracing).

Blockchain / Web3

Solidity Rust DeFi

Solidity, Rust, On-chain indexing, Wallet Auth, Settlement pipelines, DeFi risk engines.

Hire Me

I work on complex, high-performance systems across Generative AI, Computer Vision, Algorithmic Trading, Fintech/DeFi, Distributed Microservices, Full-Stack products, and Real-time platforms.

My strongest fit is work that needs hands-on engineering plus architecture: event-driven systems, CQRS/Event Sourcing, DDD, low-latency services, high-concurrency JVM systems, modern UI architecture, and Python-based machine learning pipelines.

Email Website GitHub Twitter

Pinned Loading

  1. snake game state snake game state
    1
    import { makeAutoObservable } from 'mobx';
    2
    
                  
    3
    class SnakeGameStore {
    4
      // Defining the game grid size
    5
      gridSize = 10;
  2. tmux_cheatsheet.markdown tmux_cheatsheet.markdown
    1
    # tmux cheatsheet
    2
    
                  
    3
    As configured in [my dotfiles](https://github.com/henrik/dotfiles/blob/master/tmux.conf).
    4
    
                  
    5
    start new:
  3. Demonstration of implementation "inh... Demonstration of implementation "inheritance" in clojure
    1
    ;; Define a "base type" of Dog
    2
    (defrecord Dog [breed])
    3
    
                  
    4
    ;; Define a "sub type" of TrainedDog
    5
    (defrecord TrainedDog [dog word])
  4. pole balancing benchmark for genetic... pole balancing benchmark for genetic algorithm
    1
    const MAX_TIMESTEPS = 1000;
    2
    
                  
    3
    function initializeCartPoleEnvironment() {
    4
      const gravity = 9.8; // Acceleration due to gravity, m/s^2
    5
      const cartMass = 1.0; // Mass of the cart
  5. 0. description.md 0. description.md
    1
    - `defprotocol`: defines an interface
    2
    - `deftype`: create a bare-bones object which implements a protocol
    3
    - `defrecord`: creates an immutable persistent map which implements a protocol
    4
    
                  
    5
    Typically you'll use `defrecord` (or even a basic `map`);  
  6. compatibilityDistance.js compatibilityDistance.js
    1
    function compatibilityDistance(network1, network2, c1 = 1, c2 = 1, c3 = 0.4) {
    2
      const i1 = new Set(network1.connections.map((c) => c.innovation));
    3
      const i2 = new Set(network2.connections.map((c) => c.innovation));
    4
    
                  
    5
      const disjoint =