Skip to content
View YashShelar007's full-sized avatar

Highlights

  • Pro

Organizations

@visa-lab @CSE330-OS

Block or report YashShelar007

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
YashShelar007/README.md

πŸ‘‹ Hi, I'm Yash Shelar

Software Engineer (Cloud Infra / AI-ML) @ Walnutech PBC | MSCS @ ASU

profile views


πŸš€ What I’m Working On

  • πŸ’Ό Walnutech – Building secure AWS multi-account infra, deploying microservices, and scaling AI/ML pipelines.
  • πŸ”¬ MCaaS – Cloud-native research pipeline achieving up to 14Γ— model size reduction and 3.4Γ— faster inference.
  • βš™οΈ LLMOps – Serverless answering service on AWS with Langfuse observability & CI/CD auto-rollbacks.
  • 🌐 Side Projects: ZipRide (ride-sharing app), VoyageAI (AI-driven travel planner).

πŸ›  Tech Stack Highlights

Languages: Python, Go, JavaScript/TypeScript, C++, Java, Swift, SQL, Bash
Cloud & Infra: AWS (ECS, Lambda, Step Functions, S3, API GW, DynamoDB), GCP (GKE, Cloud Functions), Terraform, Docker, Kubernetes
Frameworks: FastAPI, Flask, React, Next.js, Node.js, Tailwind CSS
Databases: PostgreSQL, MongoDB, MySQL, Redis, Firebase
ML & MLOps: PyTorch, TensorFlow, LangChain, MLflow
Observability: Langfuse (OTel), CloudWatch, Sentry


πŸ’Ό Experience (Snapshot)

Software Engineer @ Walnutech (2025 – Present)

  • Secured AWS org with IAM Identity Center + SCP guardrails (0 admin creds).
  • Deployed ECS Fargate microservices with ALB + autoscaling.
  • Self-hosted Langfuse observability β†’ cut RCA time by 70%.
  • Tech: Terraform, AWS, GitHub Actions (OIDC), FastAPI, LangChain, OpenAI.

Cloud & ML Researcher @ VISA Lab (2024 – 2025)

  • Built Model Compression as a Service (MCaaS) serverless pipeline (100+ runs/mo).
  • Reduced model size 14Γ— with <4% accuracy loss (pruning, quantization, distillation).
  • Automated infra with Terraform + Docker (CI/CD ↓ 30β†’5 mins).

πŸ“‚ Featured Projects

  • LLMOps (Serverless LLM Service) β†’ Terraform-first AWS pipeline, 3.2s cold start, p95 alerts.
  • MCaaS (Model Compression as a Service) β†’ Modular pipeline on AWS achieving 2–14Γ— size reduction.

πŸ“Š Dev Activity

From: 08 August 2025 - To: 07 November 2025

Total Time: 124 hrs 43 mins

Python               30 hrs 24 mins  >>>>>>-------------------   24.05 %
Markdown             24 hrs 32 mins  >>>>>--------------------   19.40 %
JavaScript           20 hrs 50 mins  >>>>---------------------   16.48 %
YAML                 13 hrs 25 mins  >>>----------------------   10.62 %
Terraform            10 hrs 36 mins  >>-----------------------   08.39 %
TypeScript           8 hrs 39 mins   >>-----------------------   06.85 %
Bash                 3 hrs 56 mins   >------------------------   03.12 %
JSON                 2 hrs 27 mins   -------------------------   01.95 %
HTML                 2 hrs 20 mins   -------------------------   01.86 %
Other                1 hr 43 mins    -------------------------   01.37 %

πŸ“ˆ GitHub Stats

Yash's github activity graph


πŸ’‘ Quote of the Day

Dev Quote


πŸ›  Languages & Tools (Icons)


πŸ”— Let's Connect

Pinned Loading

  1. voyage-ai voyage-ai Public

    AI Powered Trip Planner

    JavaScript 1

  2. zip-ride zip-ride Public

    Fast, Reliable Ride Sharing App

    TypeScript 1

  3. cloud-monitoring-application cloud-monitoring-application Public

    Building and deploying a cloud native monitoring application on Kubernetes

    Python 2

  4. microservices-gke microservices-gke Public

    Containerized Microservices on Google Kubernetes Engine

    Python 1

  5. serverless-architecture serverless-architecture Public

    Google Cloud Functions

    Python 1

  6. llmops llmops Public

    Production-flavored LLM answering service

    Python