Apoorv Sharma*

I'm a machine learning engineer who lives in the gap between research and reality — where a promising paper becomes a model that survives real traffic, real edge cases and real 3 a.m. pager alerts.

Over the last six years I've built vision systems that inspect factory lines, speech models that run on-device in 12 Indian languages, and retrieval pipelines that make enterprise search actually find things. My happy place is squeezing latency out of inference and watching a loss curve finally behave.

When I'm not training models, I'm writing about ML systems, mentoring early-career engineers, and over-engineering my espresso routine.

$ whoami → curious · rigorous · shipped

# apoorv@ml-rig:~/projects$ python train.py
import torch
from life import Curiosity, Coffee

model = ApoorvSharma(
    role="ML Engineer",
    focus=["vision", "speech", "llms"],
    coffee=Coffee(level=0.97),
)
model.fit(hard_problems)  # converging…