Available for workKathmandu, Nepal

Iengineerintelligentsystemsthatship.

Cloud infrastructure, machine learning in production, and full-stack product engineering. Designed carefully, shipped reliably.

AWS Certified

21

Deep-dive articles

6+

Projects shipped

AWSDockerKubernetesTerraformCI/CDLinuxPythonSageMakerMLOpsPyTorchLLM IntegrationAirflowTypeScriptReactNode.jsSQLSystem DesignMERNAWSDockerKubernetesTerraformCI/CDLinuxPythonSageMakerMLOpsPyTorchLLM IntegrationAirflowTypeScriptReactNode.jsSQLSystem DesignMERN

01 / What I do

Three disciplines, one goal: systems that work.

I move across the stack: from the infrastructure a product runs on, to the models that make it smart, to the interfaces people touch.

Cloud Engineering

Certified AWS architecture: resilient infrastructure, containers, Kubernetes, and infrastructure as code that scales without surprises.

AI & Machine Learning

From training pipelines to production inference. MLOps on SageMaker, LLM-powered features, and data workflows orchestrated with Airflow.

Software Engineering

Full-stack product development with TypeScript, React, Node, and SQL. Clean architecture, CI/CD, and observability from the first commit.

02 / Selected work

Projects I've shipped.

A snapshot of things I've designed, built, and deployed end to end.

InfraGen

AI-powered infrastructure-as-code platform: describe your infrastructure in natural language and get production-ready Terraform, architecture diagrams, and cost estimates instantly. Full-stack build with a FastAPI backend, React frontend, MinIO object storage, and one-command Docker Compose deployment.

TypeScriptFastAPITerraformLLMDocker

Spotify MLOps Recommender

Production-grade music recommendation system over Spotify's 114k-track dataset: content-based filtering with KNN and cosine similarity, real-time personalization via tunable 'sonic sliders', and zero-downtime blue/green model swapping with a user-feedback retraining loop. FastAPI + scikit-learn backend, React UI, containerized for AWS ECR/ECS.

Pythonscikit-learnFastAPIMLOpsAWSDocker

SafeTrip

2nd runner-up at the Data for Good Nepal hackathon, organized by IOA and Softwarica. Led the team from problem framing to a working data-driven prototype under hackathon deadlines.

HackathonData for GoodTeam Lead2nd Runner-Up

Anomaly Detector on AWS

End-to-end anomaly detection pipeline on AWS, provisioned entirely with Terraform as infrastructure-as-code, with reproducible environments from a single apply.

AWSTerraformPythonIaC

CSniff

Zero-dependency terminal packet sniffer in pure C: raw-socket capture, TCP/UDP/ICMP/HTTP/DNS decoding, session tracking, Wireshark-compatible PCAP logging, and port-scan/flood anomaly detection. No Npcap, no external libraries.

CNetworkingWinsock2Security

HangoutPro

Multi-sided Android platform connecting event organizers, talents, and attendees: role-based dashboards, talent booking, stories, and event logistics, built on modern Android practices with Firebase Auth.

KotlinAndroidFirebase

03 / Writing

Latest from the blog.

Long-form, hands-on guides on cloud, ML, and software engineering.

Available for opportunities

Have something worth building?

I'm open to roles and collaborations in cloud, AI/ML, and software engineering. Tell me what you're working on.