<hello/> _

Agnieszka
Mikołajczyk-Bareła

PhD · Senior AI Engineer · Researcher

Senior AI Engineer working on Reedy at Chaptr.AI — LLMs, RAG, and multimodal pipelines. Previously part of the NLP team at VoiceLab.AI that shipped TRURL, Poland's first large-scale generative model. Researcher, speaker, and co-organizer of AI-for-Good initiatives.

3 200+ citations across 30+ papers and 10+ open-source projects. Named among Top Women in AI Poland.

Skills

LLMs & AI Agents — LLM Training · Fine-tuning · RAG · LangChain · Agentic Pipelines · Prompt Engineering · Evaluation · Pydantic · OpenAI / Claude / Gemini / Llama

NLP & Speech — Transformers · BERT · GPT · TTS · ASR · Sentiment Analysis · NER

Computer Vision & Multimodal — EfficientDet · DETR · Mask R-CNN · GANs · Vision Transformers · Image Generation · Video Generation

Explainable AI — Attention Maps · Counterfactual Analysis · Bias Detection · Global Explanations

Data & Annotation — Dataset Curation · Annotation Pipelines · Data Cleaning · Benchmarking · Crowdsourcing

ML Engineering — Python · PyTorch · HuggingFace · ONNX · Model Training & Deployment

Projects

Research, open-source, and industry work

featured

Schema-First Prompting [Claude Skill]

LLM Prompt Engineering Open Source

Modern prompt engineering, designed the way a human would — clean, minimal, elegant. An open-source skill for Claude and Cursor that encodes best practices for structured outputs with Pydantic models.

Page → GitHub → Install →

npx skills add AgaMiko/schema-first-prompting

featured

The Quiet Hours [Game]

Three.js Web Game JavaScript Vibe Jam

Vibe Jam 2026: I shipped a browser 3D game—you’re the cat, she’s blocked, fetch objects and watch her novel grow. Three.js, vanilla JS, cozy until the building hints otherwise.

Page → Play →

featured

Reedy — AI Metadata Platform for Publishers

LLM RAG NLP Multimodal

Working with the team on Reedy at Chaptr.AI — an AI engine that generates keywords, THEMA codes, and SEO descriptions for book catalogs.

Page → Reedy → Chaptr →

featured

Model training

Deep Learning NLP Computer Vision Audio Open Source

Trained models across computer vision (classification, detection, segmentation), NLP and generative text, and audio; with the VoiceLab team shipped TRURL 7B/13B (incl. 8-bit and academic), vlt5-base-keywords, herbert-base-cased-sentiment, and datasets on Hugging Face.

Page → VoiceLab on Hugging Face →

featured

Data Augmentation

Deep Learning Datasets Bias Open Source

Curated review of augmentation techniques, libraries, and papers (1.6k+ stars), plus research code for Targeted Data Augmentation (TDA) and Counterfactual Bias Insertion (CBI) on dermoscopy and face data.

Page → Review ★1.6k → TDA code → Open data →

featured

Biomedical imaging

Deep Learning Medicine XAI Research

Dermoscopy / melanoma bibliography (classical features through NAS, self-supervision, augmentation, bias) plus deep learning for other biomedical imaging—microbleeds, blood smear, erythrocytes—with links added over time.

Page → GAN debiasing code →

featured

Bias in Machine Learning

Bias XAI Research Medicine Deep Learning

PhD thesis on arXiv (bias, XAI, skin-lesion case studies, style transfer / targeted augmentation / attribution feedback); IEEE Access survey with M. Grochowski; GEBI; MICCAI 2022 GAN debiasing in dermoscopy.

Page → PhD thesis (arXiv) → Survey (IEEE) → GEBI code → GEBI paper → GAN debiasing → MICCAI paper →

Waste detection

Computer Vision AI4Good Deep Learning Datasets Hackathon Open Source

WiMLDS Trójmiasto AI4Good: PyTorch detection and segmentation (EfficientDet, DETR, Mask R-CNN, Faster R-CNN) on merged litter benchmarks, Waste Management paper, curated waste image datasets list (338+ stars), and co-organized Hack4Environment (waste & environmental literacy with DIH4.AI).

Page → Detect code ★220 → Datasets review ★338 → Website → Paper → Hack4Environment →