Karanbir Singh
415 Mission Street
San Francisco, CA 94105
I am an AI/ML Engineer and Engineering Leader with 8+ years of experience building and scaling production-grade machine learning systems, agentic AI solutions, and distributed cloud platforms across fintech, automotive, and enterprise SaaS. I currently serve as a Tech Lead at Salesforce, where I lead mission-critical initiatives in cloud substrate automation, capacity and quota management, and AI-powered observability for globally distributed infrastructure with 99.99% availability.
At Salesforce, I am the product lead for automated capacity and quota management across AWS and GCP, designing decision engines that dynamically govern resource allocation based on demand, historical usage, and operational constraints. I have also led the development of agentic AI systems that automate on-call workflows, resolve infrastructure patching issues, and diagnose cloud build and provisioning failures during global data-center bring-up, reducing mean time to resolution (MTTR) by ~30%.
My research interests include Responsible AI, Retrieval-Augmented Generation (RAG), and Agentic AI, with peer-reviewed publications at ACM WWW, KDD workshops, and IEEE conferences. I am a co-author of technical books published by Springer/Apress and BPB Publications, and I regularly serve as a judge and peer reviewer for leading international conferences, journals, and industry awards. I am also an invited speaker at global AI and developer forums.
My work sits at the intersection of AI systems, cloud infrastructure, and operational excellence, with a focus on building trustworthy, scalable, and impactful AI solutions in real-world production environments.
selected publications
- Bias-Aware Agent: Enhancing Fairness in AI-Driven Knowledge RetrievalIn Companion Proceedings of the ACM Web Conference 2025 (WWW Companion ’25), 2025Presented April 28–29, 2025
- Bias Mitigation Agent: Optimizing Source Selection for Fair and Balanced Knowledge RetrievalIn Proceedings of the KDD 2025 Workshop on Agent-based Information Retrieval (Agent4IR), 2025Workshop paper at KDD 2025
- Efficient Resource Management of Kubernetes Pods Using Artificial IntelligenceIn Proceedings of the 2024 Eighth International Conference on Parallel, Distributed, and Grid Computing (PDGC), Dec 2024Presented December 19, 2024