Mukundan Madhavan

About Mukundan Madhavan

Machine learning engineer with strong expertise in both classical machine learning and deep learning. Proven track record of delivering ML-based solutions in multiple domains-I currently work as a ML lead in Google Search. In this role, I have led the construction of multiple state-of-the-art entity embeddings for the billions of entities present in Google\'s Knowledge graph. My projects have typically centred on using LLMs and LLM-based dual encoders under different architectures.Prior to this, I\'ve worked ~5-year stints in Goldman Sachs and IBM Research India. At Goldman Sachs (Technology division), my work involved predictive modelling on infra-generated datasets (telemetry generated by networks/disks, application logs etc.). As part of this, I led the delivery of highly scalable capacity outage prediction frameworks, real-time anomaly detection systems and application log-mining systems. At IBM Research, I worked on various areas including image classification and bandwidth-aware scheduling for cellular traffic.I typically lead projects through their complete lifecycle from problem formulation, ML solution design to prototyping and productionisation. I also have a strong track record of patenting and publishing my work in peer-reviewed venues.

Experience

1. Software Engineer

   Google

   Oct 2021 — Present · Bengaluru, IN

Education

• S.R.M University

  B.Tech, Electronics and Communications Engineering

  2003 — 2007

• Indian Institute of Technology, Madras

  M.S, Telecommunications

  2007 — 2010

Skills

• Algorithms
• Embedded Systems
• Android
• C++
• C
• Programming
• Matlab
• Machine Learning
• Research
• Linux
• Tcp/Ip
• Software Development
• Python