Building hardware, software, and sound
systems that matter.
HomeResumeProjectsKRX MusicAboutContact/Links
HomeResumeProjectsKRX MusicAboutContact/Links
HomeResumeProjectsKRX MusicAboutContact/Links
HomeResumeProjectsKRX MusicAboutContact/Links
HomeResumeProjectsKRX MusicAboutContact/Links
HomeResumeProjectsKRX MusicAboutContact/Links
01 / About
Building where hardware, software, and signal meet.
I’m Kroix Jones, an Electrical Engineering student at Howard University building systems
across RTL, embedded hardware, backend infrastructure, machine learning, and music technology.
Right now I’m a Hardware Systems Engineering Intern at Abbott, working near medical-device
PCBA and ASIC workflows while writing Go backend services and Python automation for
large-scale sensor analysis.
Current
Hardware Systems Engineering Intern · Abbott
Education
Howard University · B.S. EE, Math minor · GPA 3.77 · May 2028
Interests
RTL design · ASIC verification · ML systems · embedded hardware · music technology
03 / Projects
Systems that sense, compute, and adapt.
04 / Ophanim
KRX Music
A research-driven music recommendation engine built to understand taste beyond genre.
KRX Music combines perceptual audio analysis, SBERT embeddings, vector similarity,
metadata-aware ranking, and Taste Neighborhoods to model how people actually discover music.
It is my core product project and the clearest expression of how I think about ML, systems,
and sound.
01 / Audio Signal
Moving waveform enters KRX
A song starts as pressure-wave motion, then becomes structured signal behavior.
Generated Recommendation
Glass Circuit
NOVA LANE · spectral alt-R&B
Fits your catalog context with warm vocals and left-field texture.
Wave
Hash
Vector
Fusion
Pool
Card
Rank
Learn
Graph
Build
4 years
Embedding Space
384 dimensions
Alpha
100% relevance agreement
Signal
92% strongly agreed
01 / Normalize + Analyze
Waveform becomes time-frequency structure
02 / Perceptual Hash
Signal behavior compresses into binary-like fingerprints
03 / Fusion Vector
Audio, lyrics, metadata, and context become one song identity
04 / Rank + Explain
Relevance, novelty, confidence, and diversity choose the result
05 / Feedback Loop
Likes and skips reshape user-conditioned Taste Neighborhoods