Projects
Systems I’ve Built & Explored
These projects reflect my interest in cloud systems, automation, performance, and how infrastructure behaves under uncertainty, scale, and real-world constraints.
Resurge-Net — Risk-Aware Cloud Autoscaling
A research-oriented cloud autoscaling framework designed to proactively scale infrastructure by modeling demand, uncertainty, and rare workload spikes rather than reacting after SLA violations occur.
- • Built deep demand predictors using CNN and LSTM/GRU models
- • Trained on Alibaba Cluster Trace v2018 workload dataset
- • Modeled mean demand, uncertainty, and spike probability
- • Designed tunable risk-based scaling policies
- • Demonstrated improved cost vs reliability trade-offs
Python • AWS • Time-Series Forecasting • ML Systems
DocVault — Secure Serverless Document Vault
A fully serverless document management system built to eliminate server maintenance while ensuring secure authentication, controlled access, and reproducible infrastructure deployments.
- • AWS Lambda, API Gateway, S3, DynamoDB
- • JWT-based authentication and authorization
- • Presigned S3 uploads and secure file access
- • Terraform-managed infrastructure
- • GitLab CI/CD for Lambda, frontend, and IaC deployment
Custom CPU Design (Logisim)
A small CPU designed using Logisim to strengthen understanding of digital logic, datapaths, control units, and hardware-software interaction at the instruction level.
- • Datapath and control unit implementation
- • Arithmetic and logical instruction execution
- • Instruction sequencing and execution flow