Hi, I'm Lokesh.
I'm a Dublin-based Senior MLOps Engineer who enjoys building the infrastructure behind production machine learning.
My work spans distributed systems, ML platforms, and production AI infrastructure. Over more than a decade in software engineering, I've worked on distributed execution engines, reusable machine learning frameworks, and inference systems that process multimodal content.
Currently, at Integral Ad Science, I work on ML infrastructure, GPU-optimized multimodal classification, and active learning workflows. Previously, I worked on distributed execution systems at Blackswan Technologies and an internal ML framework at Noodle.ai.
This site is where I share engineering lessons, architecture trade-offs, practical examples, and experiments in inference, distributed systems, MLOps, and LLMOps. The writing reflects my own views; examples are generalized and don't disclose proprietary systems.