Arjun Bhargava

Background

About Arjun Bhargava

Senior Machine Learning Engineer with 9+ years of experience in building, optimizing, and deploying Transformer-based NLP systems, anomaly detection frameworks, and on-device ML pipelines for large-scale enterprise and cybersecurity platforms. At Samsung Research, I led AI security initiatives such as Knox Phishing Protection (KPP) and Zero Trust Endpoint Resilience (ZTER) — fine-tuning BERT/DistilBERT with PEFT and quantization for real-time on-device inference, achieving 97 % precision with < 150 ms latency across 10 M + devices. I also designed tokenizer-aware preprocessing, integrated ML inference with SoC telemetry, and developed anomaly detection from SEAndroid logs to enhance endpoint resilience. Earlier, as part of Advanced Technology Labs (ATL), I delivered ML prototypes in personalization and vision (Smart Collage, Destination Prediction, Video Summarization) and trained over 100 engineers to accelerate AI-first adoption within flagship product teams. Core Technical Expertise: – NLP & Transformers: BERT, DistilBERT, LoRA, PEFT, Tokenization Design, Transfer Learning – Model Optimization: Quantization, Pruning, Embedded Inference, Edge ML – Cybersecurity AI: SEAndroid Policy Automation, Anomaly Detection, Zero Trust Architecture – Tech Stack: Python, PyTorch, TensorFlow, C++, AWS, Git, CI/CD Passionate about advancing AI for Security and Edge Computing, I focus on transforming cutting-edge research into production-grade, privacy-preserving intelligence that runs efficiently at scale.