ML Intern @ Magic Hour (Feb 2026 - Present)

joined Magic Hour after my stint at Dria — same kind of ML/inference problems, but now across a much wider surface area.

mostly working on optimizing image, video, and TTS models for real production traffic — latency, cost, and quality tradeoffs you feel when users are generating at scale.

also spending time prototyping: new model pipelines and product-facing AI features before they get hardened for the main app.

Still ongoing — will update this post as the internship progresses.