Entry

Score

Action

Category / Wining Level

Final Score

Validated Score

Judging Score

Published

Information Technology - Cloud Computing (NEW)
(Platinum Winner)

PO#

#TBA04578

Payment Status

Paid

PO Created Date

2025-03-13 15:10:59

Paid Date

2025-03-13 15:11:58

Competition

2025

Season

Entry ID

TBE104960

Entry Submission Date

2025-03-08 08:17:53

Member's Name

Shreya Gupta

Member's Email

reachshreyagupta1@gmail.com

Entry Level

Others

Results Publish

Published

Private Entry

Yes

Hide Entry

Yes

Review

Entry Status

Category

View Entry Info

IT25024 - Cloud Computing (NEW)

Title

Company

Entrepreneur/ Executive’s Name

Submission Type

Country

Client

Linkedin Profile Link

Project Description

Project Description for Judges

Question 1 : Provide a brief description / background of your nominated organizational, departmental, team, or individual circumstance. (Your response to this question will be used for public display if your entry is awarded.)

Shreya Gupta serves as a Senior Software Engineer at a leading tech enterprise where she spearheads critical infrastructure initiatives. With 7 years of hands-on experience in cloud systems and a Master's degree in Computer Science, Shreya has built a solid foundation in sophisticated resource management. She leads the development of advanced scheduling systems for machine learning workloads, making computing resources more accessible while reducing environmental impact.
Her work on the ML Capacity Unification initiative represents a practical approach to unifying the scheduling of Machine Learning workloads across diverse computing resources. This initiative addresses the challenge of optimizing AI infrastructure by integrating two distinct workload categories to enhance fleet efficiency.
As the technical lead for the core unified scheduling layer, she created a control plane that enables secure sharing of capacity pools under defined isolation policies. Her contributions include:
Development of a capacity management layer for cloud resources
Implementation of job management strategies ensuring efficient execution
Coordination with cross-functional teams for successful ML capacity onboarding
Background
Shreya specializes in cluster management systems, advancing technologies that have shaped industry standards for resource orchestration. Her practical solutions build upon established principles while introducing sensible approaches to resource optimization.
Her work has delivered tangible results:
Optimized 5% of underutilized hardware, saving $250 million initially, with projected savings of $5 Billion over five years
Reduced 27,500 metric tons of CO2 emissions, with future reductions equivalent to removing 36,000 gasoline-powered vehicles annually
Enhanced fleet efficiency by reducing idle H100 GPUs by 15%
Improved reliability and reduced error rates in resource allocation by 8%
Beyond technical achievements, Shreya's work demonstrates her commitment to sustainable AI development. By implementing practical scheduling techniques, she has significantly reduced the carbon footprint of large-scale AI models while expanding access to AI resources by breaking down resource silos. Her approach exemplifies the balance of technical quality and environmental responsibility in modern cloud computing.

Question 2 : Outline your nominated organizational, departmental, team, or individual achievements, dated 1-3 years prior, based on your selected categories.

Technical Leadership
Shreya Gupta has established herself as a skilled technical leader in cloud resource management. With her expertise in distributed systems and scheduling algorithms, she designed and implemented the core components of the ML Capacity Unification framework. As the primary architect of the unified scheduling layer, Shreya created a system that intelligently allocates computing resources between different workload types while maintaining necessary isolation between computing environments.
Her technical contributions went beyond initial implementation as she continuously refined the scheduling algorithms based on real-world performance data. Shreya's deep understanding of resource optimization allowed her to identify previously overlooked opportunities for capacity sharing that many experts had considered impractical. Her problem-solving approach helped bridge the gap between External Workload and Internal Workload categories.
Resource Optimization
Shreya's practical approach to resource management transformed how ML infrastructure operates. By analyzing usage patterns and applying her knowledge of workload behaviors, she identified opportunities to share previously isolated resources. Through careful planning and implementation, Shreya personally optimized 5% of underutilized hardware, generating $250 million in immediate cost savings, with projected cumulative savings of $5 Billion over five years.
Her work specifically targeted the efficient utilization of H100 GPUs, which are critical components in machine learning infrastructure. By implementing sophisticated scheduling algorithms, she ensured these resources remained productive rather than sitting idle, delivering substantial financial and environmental benefits.
Environmental Impact
Shreya incorporated sustainability as core principle in her technical designs. Her scheduling techniques significantly reduced the energy usage of large ML workloads by minimizing idle time on energy-intensive hardware. This optimization work reduced 27,500 metric tons of CO2 emissions initially, with future projections equivalent to removing 36,000 gasoline-powered vehicles annually from the road.

Performance Improvements
Shreya's technical achievements extended beyond cost and environmental metrics:
Enhanced fleet efficiency by reducing idle H100 GPUs by 15%
Improved reliability and reduced error rates in resource allocation by 8%
Strengthened security measures, leading to a 10% reduction in security vulnerabilities
Increased overall system responsiveness by 20%
These performance improvements demonstrate Her ability to deliver practical solutions that address multiple technical challenges.

Question 3 : Provide examples of the strategies / plans you have introduced, in order to attain your nominated organizational, departmental, team, or individual achievements. Explain why the highlighted achievements are significant / unique.

Unified Control Plane Architecture
Shreya Gupta created a unified control plane architecture for the ML Capacity Unification initiative, enabling communication between previously isolated systems. Her design features a lightweight communication layer that maintains separation where needed while allowing resource sharing. This technical approach includes isolation mechanisms that ensure workloads remain secure while still benefiting from shared capacity pools.
The architecture incorporates event-driven communication patterns that allow real-time resource allocation decisions without creating excessive system overhead. Shreya's technical implementation maintains backward compatibility with existing systems while enabling new capabilities.
Multi-Tiered Resource Management
Shreya developed a multi-tiered approach to capacity management that functions across platform boundaries. Her system maintains continuous communication between workload management systems through a carefully designed API that handles resource requests, status updates, and deprovisioning events. This strategy allows intelligent resource sharing while respecting the operational requirements of each platform.
The resource management tier maintains a global view of capacity availability while respecting local constraints, enabling improved utilization without compromising performance or security. Shreya's implementation includes fallback mechanisms that ensure reliability even during partial system outages.

Significance and Uniqueness
What makes Shreya's work particularly significant is its practical approach to solving two critical challenges simultaneously: resource efficiency and environmental impact. Unlike previous solutions that optimized for only one metric at a time, her unified scheduling system balances multiple competing priorities.
The technical architecture Shreya designed is unique in its ability to bridge previously incompatible systems without requiring a complete redesign of either one. This pragmatic integration strategy delivered immediate benefits while setting up a foundation for future improvements.
Her work stands out because it addressed real-world constraints rather than theoretical ideals. The security model she implemented maintains necessary isolation while still allowing resource sharing, solving a technical challenge that many considered impossible. By focusing on practical implementation details and system interfaces, Shreya created a solution that works reliably in production environments where previous attempts had failed.

Optional Context for Judges (Not for publication. This section is solely for judges' evaluation.)

Thumbnail

Cancel

Link List

Link	Password Protected	Display (Winner Info)
https://en.wikipedia.org/wiki/Borg_(cluster_manager)	No	Display
https://kubernetes.io/blog/2015/04/borg-predecessor-to-kubernetes/	No	Display
https://kubernetes.io/	No	Display
https://www.forbes.com/councils/forbestechcouncil/2024/08/29/kubernetes-continuous-delivery-at-scale-business-lessons-learned/	No	Display

Credit List

Actions	Name	Title Organization

File List

Remove	Actions	File Link	File Size	File Link (System Resize)	File Size (System Resize)	Display(Winner Page)	Display Check
		upload/entry/files/TBE104960/1745399894file_9082.jpg	157.14KB	upload/entry/files/TBE104960/1745399894file_9082.jpg	157.14KB	Display

Activity Logs

Action	Details	Performed By	IP Address	Date

Edit Entry

Score

Link List

Credit List

File List

Activity Logs