LLMs + Prompting
Prompt design, model behavior, structured outputs, and practical LLM workflows.
AI Engineering / LLMs / Embedded ML / Backend Systems
Jason Kwiatkowski
I build intelligent systems with real engineering constraints: privacy-aware image infrastructure, LLM workflows, prompt engineering, embedded motion inference, algorithmic web platforms, and practical AI software.
2024
Started coding
3.7
Current GPA
7
Projects linked
AI
Engineering focus
About
Building useful AI from model behavior down to system design.
I am an AI Engineering student at Neumont College of Computer Science focused on machine learning systems, prompt engineering, LLM workflows, backend infrastructure, embedded AI, computer vision pipelines, and production-oriented software design.
My recent work connects FastAPI services, PostgreSQL and pgvector, MinIO object storage, OpenCLIP metadata, LLM prompting patterns, firmware-level inference, and algorithm visualization.
The common thread is practical intelligence: systems that stay understandable, testable, and useful when they leave the demo stage.
Embedded AI
Motion classification and firmware constraints on small devices.
Backend Infra
APIs, auth, storage, databases, Docker, and deployable services.
Privacy-Aware AI
User isolation, retention-aware storage, and careful handling of derived data.
Projects
Selected work, grouped by the year I was working on it.
Current work
2026
Neuthek
FastAPI-based image storage service combining authenticated uploads, MinIO object storage, content-aware compression, OpenCLIP metadata, and pgvector semantic search.
- Engineering value
- Production-style backend architecture with migrations, tests, privacy docs, and local infrastructure.
- Focus
- Image upload flow, vector search, metadata pipelines, retention-aware storage, and service boundaries.
FastAPI
PostgreSQL
pgvector
OpenCLIP
MinIO
Docker
View repository
Flipper Zero Drop Detector
Motion classification system for Flipper Zero using IMU feature windows, offline Python training, and compressed C-based decision-tree inference.
C
Python
Embedded ML
Decision Trees
View repository
Enigma
Class project exploring procedural maze generation, AI diffusion pipelines, and full-stack web architecture with algorithm-focused visual engineering.
C#
Algorithms
Maze Generation
Full Stack
View repository
AIE300
Full-stack item manager API using FastAPI, MongoDB, Docker, and supporting model/data experiments for applied AI engineering coursework.
FastAPI
MongoDB
Docker
PyTorch
View repository
Voice-Controlled PC Mod
Pinned GitHub project for voice-driven computer control across Windows and Linux, including volume, brightness, media, window, browser, and custom command workflows.
Python
PyQt
Speech Commands
Automation
View repository
Foundation work
2025
Starting point
2024
First learned coding
The beginning of the timeline: learning fundamentals, building early projects, and moving from curiosity into AI engineering.
Skills
Tools and systems I use to ship practical AI work.
AI + LLM Systems
Completed prompt engineering coursework and applied LLMs in practical workflows.
Prompt engineering82%
LLM workflows78%
ML systems74%
Feature engineering72%
Computer vision69%
Semantic search73%
Software Engineering
Backend development78%
Python81%
Embedded systems66%
Databases + infra74%
Python
FastAPI
PostgreSQL
pgvector
OpenCLIP
LLMs
Prompt Engineering
Docker
Redis
MinIO
SQLAlchemy
PyTorch
NumPy
scikit-learn
C
Git
Linux
APIs
PyQt
Speech Automation
Contact
Have a project, role, or AI systems problem worth building?
I am interested in AI engineering, backend systems, embedded ML, intelligent infrastructure, LLM workflows, prompt engineering, and production-oriented software projects.