Srinivasan Parthasarathy
This wiki documents the research and projects of [[people/Srinivasan Parthasarathy]], a computer science researcher at the IBM Watson Research Center who leads the Automated Machine Learning and Data Science team. His work spans cloud microservice optimization, automated machine learning, data exploration and analysis systems, interactive outlier detection, distributed systems, and distributed LLM inference. He also serves as an IBM technical lead on a collaborative project with the University of Illinois on a scalable agentic platform for multi-agent AI cloud management. He holds more than twenty granted US patents and has published over 100 papers across theoretical computer science, algorithms, data mining, and networking. For a full biography, see [[people/Srinivasan Parthasarathy]].
Projects
- [[projects/AgraBOT]] — RAG framework for third-party security risk management
- [[projects/blis]] — blackbox inference simulator for LLM inference calibration
- [[projects/Cognito]] — automated feature engineering for supervised learning
- [[projects/Foresight]] — visual insight recommendation engine
- [[projects/hybrid-cloud-ai]] — agentic platform for multi-agent AI cloud management (IBM technical lead)
- [[projects/Iter8]] — Kubernetes-native online experimentation and canary releases
- [[projects/llm-d]] — distributed inference framework for LLMs on Kubernetes
- [[projects/READ]] — rapid data exploration, analysis and discovery
- [[projects/REMIX]] — interactive outlier detection system (KDD 2017)
- [[projects/SOFIA]] — ontology-driven analytics recommendation system (ISWC 2015)
- [[projects/TraceWeaver]] — distributed request tracing for microservices
Patents
- [[patents/patent-portfolio|Srinivasan Parthasarathy Patent Portfolio]] — complete listing of more than twenty granted US patents
- [[patents/Application Topology Discovery]] — automated topology discovery using sandbox analysis and passive production queries
- [[patents/shiftleft-topology-construction|Shiftleft Topology Construction and Information Augmentation Using Machine Learning]] — shiftleft microservice topology construction using source code analysis and graph convolution networks
- [[patents/auto-tuner-cloud-micro-services-embeddings|Auto Tuner for Cloud Micro Services Embeddings]] — ML-driven auto-tuning of microservice timeout and retry parameters
- [[patents/fault-localization-alert-aggregation|Fault Localization and Alert Aggregation]] — root cause analysis using topological sort on directed graphs
- [[patents/multi-objective-optimization-of-applications|Multi Objective Optimization of Applications]] — Pareto surface optimization for multi-KPI application tuning