Skip to content

DEV Community

Arsen Apostolov

Head of AI & Process Transformation @ Postbank | RPA, GenAI & local-first AI | Building homelab-monitor & patzer | Find me on github.com/SikamikanikoBG & medium.com/@arsen.apostolov

Bulgaria Joined on Apr 7, 2024 https://medium.com/@arsen.apostolov

ImageGenerator

Generate images using HuggingFace or local models based on Stable Diffusion. Compare generated images from miscelaneous models using the same prompt. Batch images creation.

Python

employee_development_matrics

Python

Python, AI/ML, Data Science, Business Analytics, CRM
Specializing in:
- Generative AI implementation
- Business strategy with AI
- Data product development
- AI team leadership

Currently exploring generative AI tools and rapid prototyping. Completed Python fundamentals and Neural Networks certifications. Always learning new ways to leverage AI in business applications.

Building AI-powered tools for financial services, exploring LLM applications, and working on data democratization projects. Writing about AI implementation experiences and best practices.

- AI/ML consulting and discussions
- Insights about data-driven culture
- Collaborating on AI projects
- Writing about practical AI implementations in business
- Mentoring in data science and AI

32 posts published

45 comments written

0 tags followed

Arsen Apostolov

Sep 26

Qwen3.8-Flash-Next (125B) on Three RTX 3090s at 80 Tokens/s: Teaching llama.cpp Which Experts Matter

#llm #qwen #homelab #ai

15 min read

Want to connect with Arsen Apostolov?

Create an account to connect with Arsen Apostolov. You can also sign in below to proceed if you already have an account.

Already have an account? Sign in

Arsen Apostolov

Sep 23

Qwen3.8-27B on One RTX 3090 vs Two: +20% Decode, +14% Cold Prefill, and 3x on Cached Prompts

#llm #ai #qwen

7 min read

Arsen Apostolov

Sep 20

I Benchmarked Qwen3.8-27B on Three RTX 3090s. The Second Card Bought Me 13% More Decode — and a 3x Faster Prefill

#llm #qwen #homelab #ai

1 min read

Arsen Apostolov

Sep 20

What Does a Local LLM Actually Cost per Month? I Read the Meters.

#ai #hardware #llm

5 min read

Arsen Apostolov

Sep 8

Qwen3.8-27B on 2 RTX 3090: My First Local Model I Actually Trust

#llm #ai #homelab #selfhosted

3 min read

Arsen Apostolov

Jul 9

Does a Second GPU Increase Ollama's Context Window? (Quadro P2000 + RTX 3090 Tested)

#llm #ollama #vllm #gpu

3 min read

Arsen Apostolov

Jul 8

Whisper large-v3 VRAM Requirements: Why It Won't Fit on a 5GB GPU (and What We Tried Instead)

#machinelearning #whisper #gpu #homelab

5 min read

Arsen Apostolov

Jul 5

vLLM vs llama.cpp vs Ollama: What Happens When Your Model Doesn't Fit in 24GB VRAM

#llm #homelab #vllm #ai

6 min read

Arsen Apostolov

Jul 3

Local LLM vs Claude: Benchmarking qwen3-coder:30b as a Production Agent Backend

#llm #homelab #opensource #ai

5 min read

Arsen Apostolov

Jun 28

How to Run Reliable Local LLM Agents on an RTX 3090: A Benchmark (5 Models, Priced in Watts)

#llm #homelab #opensource #ai

3 min read

Arsen Apostolov

Jun 23

How to Rank Local LLMs by Cost per Correct Answer (Measured GPU Energy, 8 Ollama Models)

#machinelearning #llm #gpu #opensource

4 min read

Arsen Apostolov

Jun 22

How Much Does It Actually Cost to Run a Local LLM? (€ per Million Tokens, Measured)

#machinelearning #llm #homelab #opensource

2 min read

Arsen Apostolov

Jun 22

How to Fix Watchtower Not Updating Containers on Docker 29

#docker #selfhosted #homelab #devops

1 min read

Arsen Apostolov

Jun 21

How to Fix Docker Networking After a firewalld Reload

#docker #linux #selfhosted #homelab

1 min read

Arsen Apostolov

Jun 21

How to Get Disk-Full Alerts Across Linux and Windows

#linux #windows #sysadmin #homelab

2 min read

Arsen Apostolov

Jun 7

I want to let an AI agent roam my homelab — looking for someone to build the MCP server

#help #opensource #selfhosted #ai

2 min read

Arsen Apostolov

Jun 7

The homelab box you forgot you own is probably 47 updates behind — here’s the safe fix

#homelab #selfhosting #linux #sysadmin

3 min read

Arsen Apostolov

Jun 6

Reclaiming 16GB of idle VRAM: a 30-line sidecar that evicts ComfyUI when it stops working

#homelabselfaidockerhosted

4 min read

Arsen Apostolov

Jun 3

Fitting WhisperX large-v3 + a 24B LLM on one 3090: a reproducible context-capping recipe

#homelab #ollama #localllm #devops

5 min read

Arsen Apostolov

May 26

I got tired of guessing which model holds my VRAM, so I built a tiny dashboard

#showdev #ai #docker #monitoring

1 min read

Arsen Apostolov

Mar 22

Building a Game for My Daughter with AI — Part 1: What If She Could Build It Too?

#ai #programming #opensource #webdev

1 min read

Arsen Apostolov

Jun 7 '25

I Hit a $200 AI Bill and Built My Own Server Instead - Complete Guide!

#programming #ai #coding #remote

1 min read

Arsen Apostolov

Mar 23 '25

Stop Copy-Pasting Your Code to LLMs — I Built a Tool That Does It Automatically

#discuss #llm #productivity #tooling

1 min read

Arsen Apostolov

Feb 17 '25

Testing Aider: Practical Experience with Different Models

#coding #ai #powerfuldevs #programming

1 min read

Arsen Apostolov

Jan 10 '25

I’m sharing my LLM Code Lens Python package: My Secret Weapon for AI-Powered Coding

2 min read

Arsen Apostolov

Jan 5 '25

My Weekend project on GitHub: Making AI Art Creation Simple For Everyone 🎨

#ai #python #opensource #beginners

4 min read

Arsen Apostolov

Jan 3 '25

Looking for feedback on the next article about home lab setups - AI, automation, or ML?

#homelab #automation #learning #ai

1 min read

Arsen Apostolov

Jan 2 '25

5 Python Functions That Will Speed Up Your Data Analysis 🚀

2 min read

Arsen Apostolov

Dec 29 '24

5 Maintenance Habits That Keep Your Dev Machine Purring

#productivity #maintenance #developers

1 min read

Arsen Apostolov

Dec 26 '24

GitHub Copilot Now Free: why is it important? 🤔

2 min read

Arsen Apostolov

Dec 26 '24

GitHub Copilot Now Free: why is it important? 🤔

#ai #aistrategy #futureofwork #github

2 min read

Arsen Apostolov

Dec 25 '24

Running Out of Space? Move Your Ollama Models to a Different Drive 🚀

#ollama #ai #tutorial

1 min read

loading...