Nishikant Komerishetty Padige

Product Engineer

San Francisco, California, United States3 yrs 7 mos experience
AI EnabledAI ML Practitioner

Key Highlights

  • Spearheaded GPU cluster operations at Meta.
  • Designed real-time data processing pipeline for smart homes.
  • Developed secure chat protocol enhancing user privacy.
Stackforce AI infers this person is a highly skilled AI/HPC and Software Engineer with expertise in Machine Learning and Networking.

Contact

Skills

Core Skills

High Performance ComputingDistributed SystemsGenerative AiMachine LearningNetwork EngineeringAndroid Development

Other Skills

Computer NetworkingData StructuresData AnalysisData VisualizationAmazon Web Services (AWS)React.jsApache FlinkAWS KinesisObject-Oriented Programming (OOP)CCmockaPython (Programming Language)Cisco NetworkingAndroid StudioXMPP

Experience

3 yrs 7 mos
Total Experience
1 yr 3 mos
Average Tenure
11 mos
Current Experience

Meta

AI/HPC Production Engineer

Jul 2025Present · 11 mos · Menlo Park, CA · On-site

  • Spearheaded massive-scale cluster operations, onboarding and managing the performance of over 100,000 GPUs in total, including advanced GB200 and GB300 architectures.
  • Optimized distributed LLM training infrastructure by managing high-performance GPU cluster networking, heavily utilizing RDMA to ensure GPU communications never bottlenecked high-scale training jobs.
  • Diagnosed and resolved critical performance bugs in major training workloads, successfully increasing overall GPU communication performance by 40%.
  • Engineered a topology-aware NIC selection algorithm for GPU-initiated comms within PyTorch/torchcomms, optimizing network routing and hardware utilization (TorchComms PR #698).
  • Accelerated LLM inference pipelines by working closely with SGLang and vLLM inference engines, integrating and tuning distributed KV cache systems like Mooncake.
  • Doubled distributed KV cache efficiency, improving Mooncake performance by 100% for Meta's internal clusters, and successfully committed these system optimizations back to the open-source community (Mooncake PR #2015).
  • Re-engineered LMCache lookup mechanisms with Mooncake, optimizing the cache lookup time complexity from O(N) to O(1) and merging the enhancement upstream to open-source (LMCache PR #2976).
Computer NetworkingGenerative AIHigh Performance ComputingDistributed Systems

Arizona state university

2 roles

Research Assistant

Oct 2024May 2025 · 7 mos · Tempe, Arizona, United States

Data AnalysisData Visualization

Research Assistant

Aug 2024Nov 2024 · 3 mos · Tempe, Arizona, United States

Amazon Web Services (AWS)React.js

Progress residential®

Software Engineer

May 2024Aug 2024 · 3 mos · Tempe, Arizona, United States · On-site

  • Designed and developed a real-time data stream processing pipeline using Apache Flink in Java and AWS Kinesis to enhance intruder detection in smart homes.
  • Implemented multi-GPU parallel processing on AWS SageMaker to optimize data preprocessing and model inference, reducing inference time by 75%.
  • Developed an ensemble machine learning model combining SVM, Random Forest, and MLP classifiers to classify properties under trespasser threat with 85% accuracy.
Machine LearningAmazon Web Services (AWS)

Cisco

2 roles

Software Engineer

Promoted

Aug 2021Jul 2023 · 1 yr 11 mos

  • Designed and integrated modular test framework enhancements using C and Cmocka, incorporating modular compilation to decrease test compilation time by 93% and improve test coverage by 30%.
  • Diagnosed and resolved critical memory leaks in Layer 2 VPN configurations using heap analysis and leak detection with Valgrind, ensuring stable and reliable performance.
  • Restructured the tracing system using LTrace, standardizing trace logs, which enhanced code clarity and enabled 30% faster resolution of network issues.
  • Created Yang models for Layer 2 VPN configurations, enabling complete feature configuration through an intuitive GUI, improving the user experience and feature accessibility for network operators.
  • Awarded quarterly quality award by Cisco Systems for producing high-quality and innovative solutions
Object-Oriented Programming (OOP)Network Engineering

Software Engineer

Jan 2021Jul 2021 · 6 mos

  • Streamlined debugging and boosted code clarity by restructuring the tracing system.
  • Diagnosed and fixed memory leaks causing Layer 2 VPN tunnel instability.
Python (Programming Language)Cisco Networking

Spaarks

Android Developer

Jun 2020Dec 2020 · 6 mos · Hyderabad, Telangana, India · On-site

  • Developed a custom protocol using XMPP and Java, integrating encryption techniques to secure and anonymize user chat functionality, enhancing privacy and communication security.
  • Diagnosed and resolved a critical performance bottleneck in the application by optimizing threading and memory allocation, improving responsiveness by 50%, and significantly enhancing user experience and application efficiency.
  • Engineered and optimized the notification system by tuning the inbuilt chat system to efficiently reuse existing resources, improving system performance and responsiveness.
Object-Oriented Programming (OOP)Android StudioAndroid Development

Education

Arizona State University

Master's degree — Computer Science

Aug 2023May 2025

Jawaharlal Nehru Technological University Hyderabad (JNTUH)

Bachelor of Technology - BTech — Computer Science

Aug 2017Jun 2021

Indian School of Business

Certification course — Technology Entrepreneurship Program

Jan 2019Jan 2020

Stackforce found 100+ more professionals with High Performance Computing & Distributed Systems

Explore similar profiles based on matching skills and experience