Manish Pandey

마니쉬 판데이

Ph.D. candidate in Computer Science and Engineering. Systems research on serverless computing, GPU runtimes, and reinforcement learning for resource management.

Daegu, South Korea
manishpandeyabc@gmail.com
+82 10-2199-7965

Summary

Seven years of systems research on serverless computing, GPU runtimes, and reinforcement learning for resource management. Author of ten peer-reviewed papers, including IEEE Transactions on Cloud Computing and ACM/IFIP Middleware. Builds and evaluates production-grade infrastructure on Apache OpenWhisk, Kubernetes, and CUDA, validated against Azure and Alibaba production traces. Contributor to 14 nationally funded R&D projects spanning disaster response and industrial safety.

Research Experience

Graduate Research Assistant

Sept 2021 – Present
Intelligent Software Systems Lab (ISSLab), Kyungpook National University
  • NMIG — a GPU runtime for serverless ML inference on Apache OpenWhisk that decouples GPU contexts from container lifetimes, cutting cold-start latency and raising GPU utilization over stock OpenWhisk. Accepted to ACM/IFIP Middleware 2026.
  • D3QL / PROBA — an event-driven Double Dueling Deep Q-Learning offloading and scheduling policy for serverless edge clusters, reducing deadline misses and end-to-end latency. Published in IEEE Transactions on Cloud Computing (Q1) and IEEE CLOUD 2025.
  • FuncMem — a memory right-sizing and adaptive execution layer that predicts per-function memory footprints, cutting cold-start latency by 63.5%, memory allocation by 47.0%, and cumulative execution time by 54.9%, while raising throughput from 0.76 to 1.63 functions per second and lowering average queueing delay from 5.2 to 2.0 seconds. Published at ACM SAC 2024.
  • Implemented a priority-based, deadline-aware function scheduler and evaluated all four systems on an OpenWhisk testbed driven by Azure Functions and Alibaba production traces.
  • Released two registered software works, SECS and RLECO, in Python and Go with CUDA-backed inference paths.

Graduate Researcher

Sept 2019 – Aug 2021
Kyungpook National University
  • Built middleware for mobile edge computing that profiles device heterogeneity and places computation across edge and cloud tiers. Published at IEEE SMDS 2021 and ITC-CSCC 2021.
  • Co-developed a crowdsourced earthquake early-warning system on commodity smart devices, published at IEEE/ACM MobileSoft 2021.

Publications

Eleven peer-reviewed publications. Author name in bold; ⋆ marks papers presented in person by the author.

Journal Articles

  1. J1
    Event-Driven Double Dueling Deep Q-Learning for Efficient Serverless Edge Function Offloading and Scheduling
    Manish Pandey, Young-Woo Kwon
    IEEE Transactions on Cloud Computing (Q1). 10.1109/TCC.2026.3703387
    2026
  2. J2
    Development of an Image Processing-Based Ball Tracking Program for Table Tennis
    Siddhartha Bikram Panday, Je-heon Moon, Seung-hun Lee, Ju-sung Lee, Manish Pandey, Gustavo Adrian Ruiz Sanchez
    Korean Journal of Sport Science. 10.24985/kjss.2020.31.2.180
    2020

Conference Papers

  1. {{ p.id }}
    {{ p.title }}
    {{ p.authors }}
    {{ p.venue }}
    {{ p.year }}

Technical Reports

  1. R1
    Stabilization and Practicability Verification of a Low-cost Earthquake Motion Alarm
    Kyungpook National University research team (contributor: Manish Pandey)
    National Disaster Management Research Institute (NDMI) project report. Download report
    2020

Software and Intellectual Property

SECS 2026
Serverless Edge-Cloud Computing Simulator. Registered software copyright, Korea Copyright Commission (registered 2026.04.29; created & published 2026.03.24). Copyright holder: Kyungpook National University Industry-Academic Cooperation Foundation.
RLECO 2026
Reinforcement-learning edge-cloud orchestrator. Registered software copyright, Korea Copyright Commission.
Patent application Filed 2024
Method for Calculating Seismic Intensity Scales of Surface Using Table-Based Calculation (테이블 연산에 기반한 진도 산출 방법). Korean patent application 10-2024-0046869. 10.8080/1020240046869

Education

Ph.D. in Computer Science and Engineering

Sept 2021 – Feb 2027 (expected)
Kyungpook National University, Daegu, South Korea · GPA 3.7/4.3
Dissertation: Serverless Computing as a Substrate for Heterogeneous AI Workloads. Advisor: Prof. Young-Woo Kwon.

M.S. in Computer Science and Engineering

Sept 2019 – Sept 2021
Kyungpook National University, Daegu, South Korea · GPA 3.95/4.3

B.Sc. in Computer Science and Information Technology

Nov 2011 – Oct 2015
Tribhuvan University, Kathmandu, Nepal · 72.5/100
Korean Language Program, Kangwon National University, Samcheok (Mar 2018 – Aug 2019).

Professional Experience

Senior Software Engineer, Swift Technology

July 2017 – Mar 2018
Kathmandu, Nepal
Built and maintained transactional web applications in PHP and MySQL, and integrated the IMEPAY payment gateway into a ticketing platform with secure transaction handling.

Software Developer, Smartmobe Solutions

Apr 2016 – July 2017
Kathmandu, Nepal
Full-stack development in PHP and AngularJS on Azure, delivering a real-estate marketplace and a staff behavior analysis system.

Junior Software Developer, Makura Creations

June 2015 – Apr 2016
Kathmandu, Nepal
Developed and maintained client websites in PHP, JavaScript, and MySQL for travel, hospitality, and small-business customers.

Funded Research Projects

Student researcher on 14 government- and industry-funded projects at Kyungpook National University.

{{ pr.en }}
{{ pr.ko }}
{{ pr.period }}

Technical Skills

{{ s.label }}
{{ s.items }}

SONSIK Presidency

President of the Society of Nepalese Students in Korea (SONSIK) for the 19th Executive Committee, 2022–23 — leading a nationwide student organisation across scholarship, cultural, advocacy and publication programmes.

  • Scholarships to more than 320 students across five districts and provinces of Nepal; revived the SONSIK Journal and the annual Smarika.
  • Led 30-plus Dashain, Tihar, Teej and Chuseok cultural programmes and represented Nepal at UNESCO, MAMF and city multicultural festivals.
  • Ran webinars on the GKS scholarship, CV writing and voice-phishing safety; coordinated emergency medical fundraising and the 20th Annual General Meeting.
View the full presidency record →

Honors, Service and Languages

Winner, KNU Global Challenger Program — week-long research excursion to NCSA, Argonne National Laboratory, and UC Berkeley
2022
Peer reviewer — reviewed 11 manuscripts for international journals and conferences (verified on ORCID)
2022 – 2026
Society of Nepalese Students in Korea (SONSIK) — Advisor (2023–24), President (2022–23), General Secretary (2021–22), Member (2020–21)
2020 – 2024
Languages
Nepali native · English fluent · Hindi fluent · Korean conversational (8+ years residence, completed Korean Language Program)