guest@portfolio:~$ ./welcome.sh

Kushal Subedi_

Data Engineer & Researcher.

profile.jpeg
Kushal Subedi

// Current Status
Specializing in real-time distributed systems, Multi-Agent Systems, and blockchain analytics. PhD Student @ KSU.

git checkout projects cat cv.pdf contact_me()

01. Experience

Graduate Research Assistant

Jan 2026 - Present
@ Kennesaw State University, CCSE
  • ▹ Conducting research in Distributed Artificial Intelligence (DAI) and Multi-Agent Systems (MAS).
  • ▹ Designing multi-agent frameworks for automated software quality assurance and code review.

Data Engineer

Feb 2025 - Dec 2025
@ 108 Capital Pvt. Ltd.
  • ▹ Architected production-grade Kappa Architecture pipelines processing 10+ TB of Solana blockchain data, deploying self-hosted Apache Kafka and Apache Flink clusters with exactly-once semantics.
  • ▹ Engineered a scalable historical data ingestion pipeline using OldFaithful archive and Subsquid, optimizing a self-hosted StarRocks OLAP cluster to achieve 10x query performance improvements.
  • ▹ Built fault-tolerant real-time block streaming infrastructure leveraging gRPC and Redis Streams for deduplication, ensuring zero data loss during provider failures.
  • ▹ Developed modular SQL transformation pipelines with dbt Core, optimizing query performance through strategic materialization configurations (tables, views, and incremental models).
  • ▹ Developed high-throughput transformation layers using Bento (Benthos) with custom Go plugins to parse raw block data and extract transaction instructions.
  • ▹ Managed end-to-end cloud infrastructure using Docker Swarm and GitOps workflows; established a full observability stack with Prometheus and Grafana.

Associate MLOps Engineer

Feb 2024 - Feb 2025
@ TAI Inc.
  • ▹ Conducted in-depth research on Large Language Models (LLMs) and successfully integrated their capabilities into a document relation analyzer, enhancing semantic extraction accuracy.
  • ▹ Engineered robust Python scripts for efficient graph database management and complex relationship manipulation within the Neo4j environment.
  • ▹ Leveraged Pandas and NumPy libraries for comprehensive data analysis, troubleshooting, and optimization of complex data operations to ensure system reliability.
  • ▹ Designed and implemented automated Jenkins pipelines and custom shell scripts to streamline data engineering processes, significantly enhancing operational efficiency.
  • ▹ Developed comprehensive unit test suites using Pytest to validate data pipelines and ML models, ensuring code reliability and seamless CI/CD integration.

02. Featured Projects

Image Inpainting with U-Net

Implemented U-Net architecture with partial convolution layers on 50K+ images. Achieved 28.5 dB PSNR and deployed as a scalable RESTful API.

PyTorch OpenCV Azure

IOUMate: P2P Financial Ledger API

Engineered a robust RESTful API for a peer-to-peer debt management platform using Django REST Framework. The system features a sophisticated verification-based transaction engine, ensuring data consistency between lending and borrowingparties through a shared ledger logic.

Django REST OpenCV Azure

Videolize: Script-to-Video

Winner (GenAI Category) at Hack-A-Week 2023. Pipeline converting text to video using diffusion models (SDXL) and SLERP interpolation for smooth transitions.

SDXL Gemini API Python

Smartbot: 29 Card Game Agent

Winner of Smart Bots Coding Challenge 2023. Information Set MCTS agent handling imperfect information for 10K+ parallel simulations.

RL MCTS

03. Education

PhD in Computer Science

Kennesaw State University

Jan 2026 - Present

Bachelor's in Computer Engineering

Tribhuwan University

GPA: 3.84

2024

04. What's Next?

Get In Touch

I'm always open to discussing new opportunities, collaborations on distributed systems research, or just having a chat about AI.

Say Hello