For over five years, I’ve been part of the low-code and automation revolution, architecting sophisticated solutions that bridge the gap between powerful platforms and custom development. My expertise lies in transforming complex business processes into streamlined, intelligent, and automated systems.
I began working with OpenAI’s API in early 2021 (GPT-3 Beta era). This experience now allows me to move beyond simple prompts and engineer robust agentic automations, systems that can reason, plan, and execute complex, multi-step tasks autonomously.
I began working with OpenAI’s API in early 2021 (GPT-3 Beta era)
My day-to-day work is deeply technical. I’m not just a user of tools, I extend them.
This includes:
Integrating diverse systems through REST APIs and Webhooks.
Writing custom server-side logic using database edge functions (e.g., in Supabase).
Developing custom nodes in n8n to create bespoke functionality not available out-of-the-box.
Managing application deployment and infrastructure on platforms like Railway.
I thrive on building the connective tissue that makes modern software work, creating seamless web applications and powerful backend automation that drive real-world results.
Core things I use on a daily basis:
- n8n
- Supabase
- RestAPIs & Webhooks
- Railway
- OpenAIs developer platform
- Airtable
Skills
Language
Work Experience
Education
Qualifications
Industry Experience
The Problem: Operations required checking 100% of photos, but manual review covered <1%. Additionally, finding who performed a job required a manual search through Excel sheets.
The Solution: Built a custom Computer Vision pipeline (YOLOv12/Roboflow) that audits images and provides feedback to the worker in <20 seconds (via API and Supabase Edge Functions).
Admin Automation: Fully automated the “Worker Lookup” process, matching Job IDs to workers instantly, saving hours of daily manual Excel work.
Internal portal: Created an internal portal to showcase operations data - how many jobs are done in each
location, worker statistics, etc.
• The problem: Sales and CS teams need to check in with prospects they had demoed before, while also making new demos to other prospects
• The Solution: Build a voice agent with a “Pre-Flight” Reasoning Engine, and only prospects that fit certain criteria were called to. The AI agent had full context of the prospect and the capability to book a new call between the prospect and the sales team.
Hire Karlis today
To get started post up your job and then invite Karlis to your job.