Prateek Mehra

Applied ML Engineer building production computer vision systems

4 years experience · Bengaluru

Built and deployed production computer vision pipelines serving 100M+ images/month across batch and real-time workloads.

450+ outlets 100M+ images/month 40k+ daily inferences 0.947 mAP@0.5

Experience

Applied ML Engineer - Data Tech, EatClub Brands (Box8)

Bengaluru · 2022 — Present

  • Built and deployed production computer vision pipelines for operational auditing, anomaly detection and OCR across 450+ outlets.
  • Owned end-to-end ML lifecycle: data curation → training → deployment → monitoring → iterative threshold tuning.

Currently exploring: edge AI · real-time CV · multimodal systems · low-latency inference

Blogs

When Are You Done Speaking? Frozen Whisper Embeddings vs Handcrafted Features

Turn detection with Whisper Tiny + XGBoost: an ablation study, a scaling curve, and one tuned threshold · 5 min read

On-Device Vision at 4% CPU: FP16 Core ML vs PyTorch MPS

YOLOv8n bottle detection on an M1 Air that sleeps between frames · 5 min read

Skills

ML: PyTorch, TensorFlow, OpenCV, scikit-learn
Systems: Airflow, Docker, Linux, real-time inference
Infra: AWS, GCP, monitoring, cloud-edge deployments
Languages: Python, SQL