Full-Stack Engineer|Automation & LLM Fine-Tuning|Python|Celery|Azure

Overview

Looking for full-time work (8 hours/day)

at $12.22/hour ($2,363.65/month)

Not Specified

Last Active

July 27th, 2026 (yesterday)

Member Since

January 22nd, 2018

Profile Description

Experienced Fullstack Developer with strong backend focus. Skilled in Python, Flask, Django, Celery, Businesscentral 365, SQLAlchemy, REST APIs, and cloud platforms (AWS, Azure). Proficient in Postgres, MySQL, and JavaScript, with hands-on experience in web scraping and external integrations (Amazon SP-API, Amazon Ads API, eBay API). Experienced in building applications from the ground up to deployment, including DevOps and CI/CD workflows. Self-directed and resourceful. Currently working in automation with screenreader with generative AI and LLM technologies.

AI Systems & Fine-Tuning
Design and run LLM fine-tuning pipelines end to end: QLoRA/LoRA training (PEFT, TRL, bitsandbytes) across model families from 0.5B to 32B parameters, behind a config-driven acceptance gate that checks both task improvement and regression safety before anything ships. Built and validated three distinct evaluation approaches for a fine-tune's gain - LLM-as-judge (pairwise and decomposed criterion-based scoring), exact-match structured/classification metrics, and blind human-label calibration - and have direct experience diagnosing when an automated judge's verdict can't be trusted, including a replicated content-bias finding that held up across two different judge architectures and families. Comfortable with the full surrounding stack: dataset curation and synthetic data generation, GPU rental infrastructure (RunPod), and diffusion/image-model training and workflow engineering (AI-Toolkit, ComfyUI, AUTOMATIC1111/Forge - Stable Diffusion WebUI)