Yash Sahsani
I build AI-enabled developer platforms and reliable cloud systems across LLMOps, MCP, RAG, observability, DevOps automation, and full-stack product delivery. My current work combines RBC platform engineering, machine learning workflows, and production-grade SRE practices.
Personal operations map
AI + SRE profile
ML
Engineer 1 @ RBC
Sr
Software Developer
60%
CI/CD time reduction
GCP
ACE certified
atlas-mcp --serve prompts --rag catalog --observe otel --deploy openshift
A monorepo-inspired map of how I design engineering platforms.
The portfolio now presents my work like a living system: product surfaces, reusable packages, runtime infrastructure, and operational feedback loops all visible in one place.
/platform-workspace
Monorepo systems map
apps
packages
infra
Design engineer mode
Interfaces that explain the system
Prototype
Turn a platform problem into a usable workflow before overbuilding the system.
Instrument
Add traces, alerts, service ownership, and operational signals while the feature is still fresh.
Harden
Move from proof of concept to stable paths with templates, guardrails, and deployment checks.
Ship
Release through repeatable CI/CD and keep feedback loops close to the engineers using it.
Built around the exact systems I work on.
My work spans AI developer infrastructure, cloud platform reliability, observability, production support, and full-stack delivery across Canada and India.
AI platform engineering
LLMOps, MCP, RAG, prompt infrastructure, entity catalogs, and Backstage-powered developer workflows.
Cloud automation
Terraform, GitLab CI, GitHub Actions, Docker, Cloud Run, Cloud Functions, OpenShift, and repeatable environments.
Observability systems
Dynatrace, OpenTelemetry Collector, Splunk, PagerDuty, Kafka observability, SLIs, and SLOs.
Full-stack builds
Next.js, NestJS, Django, Flask, Redis, MongoDB, and real-time product workflows.
Security-aware engineering
Practical exposure to blockchain, reverse engineering, IAM, and infrastructure hardening.
Switch between the exact work streams shaping my current path.
LLMOps + MCP + RAG
Machine Learning Engineer 1
RBC • Toronto, ON
Built an ATLAS MCP Server in Node.js to serve AI prompts across IDE workflows.
Created a RAG proof of concept over a structured knowledge base and entity catalog.
Focused on semantic search, ownership discovery, and developer AI enablement.
From machine learning systems to operational reliability.
The throughline is ownership: design the platform, deploy the system, instrument it, learn from incidents, and turn those lessons into better developer workflows.
01
May 2026 - Present
Machine Learning Engineer 1
Royal Bank of Canada (RBC) • Toronto, ON
Building AI platform workflows around LLMOps, Model Context Protocol (MCP), RAG, and developer experience. Built an ATLAS MCP Server in Node.js to dynamically fetch and serve AI prompts across IDEs through Backstage.io APIs. Developed a proof-of-concept RAG system using a structured knowledge base and entity catalog for semantic search and ownership discovery.
02
Apr 2025 - May 2026
Senior Software Developer
Royal Bank of Canada (RBC) • Toronto, ON
Configured observability for legacy mainframe C-based applications using OpenTelemetry Collector, Dynatrace, and OpenShift. Supported Kafka observability for asynchronous notification systems and helped improve operational visibility for platform teams. Built Launchpad tooling to automate repository creation with standard Helios and Artemis DevOps configuration.
03
Sep 2024 - Apr 2025
Software Developer (Digital DevOps And Observability)
Royal Bank of Canada (RBC) • Toronto, Ontario, Canada
Configured monitoring and alerting systems using Dynatrace, Splunk, and PagerDuty, improving issue detection and response times. Centralized application insights with a Dynatrace integration on the RBC Developer Portal built on Backstage.io. Automated CI/CD pipelines with Jenkins and GitHub Actions to implement monitoring-as-code and continuous health checks.
04
June 2022 - April 2024
Cloud Engineer
Techwondoe • Hyderabad, India
Developed backend and cloud services for invoice processing with Python, TypeScript, Docker, Terraform, GitLab CI, and PostgreSQL. Deployed serverless infrastructure across DEV, SIT, UAT, and PROD on GCP using Cloud Run, Cloud Functions, and Terraform. Led the incident response team, ensuring rapid resolutions and implementing permanent fixes for production issues to maintain service reliability.
05
May 2020 - June 2020
Software Developer
Cilans System • Ahmedabad, Gujarat, India
Developed a Hyperledger-based marketplace for automobile sales, implementing Node.js chain code and leveraging Azure Cloud Services to create a private blockchain. Used Scrum and Agile practices for delivery, sprint planning, code quality, and cross-team collaboration. Automated project deployment on Azure using CI/CD YAML scripts with GitHub Actions, ensuring efficient release cycles and improved deployment reliability.
Selected builds with AI, cloud, observability, security, and product depth.
YoutubeNLP
YouTubeNLP is a microservices-based machine learning application that classifies YouTube videos using NLP techniques including sentiment analysis, emotion detection, keyword extraction, ETL pipelines, and FastAPI model inference.
CodeOps
Automated cloud application deployment on AWS EC2, ECS, S3, and Lambda with Terraform, applying Infrastructure as Code, IAM hardening, and repeatable module design for reliable environments.
DynaTrax
Dynatrax is an observability solution that integrates OpenTelemetry with Dynatrace for a Python Flask application. It enables you to capture, track, and analyze traces, metrics, and logs in real-time, giving you deep insights into your application’s health and performance.
AnisbleCI
Incorporated Ansible within a CI/CD pipeline to automate application deployments, enabling Jenkins to concentrate on artifact creation.Set up a DockerHub repository to store and manage Docker images, simplifying the deployment.Configured an Ansible server on an EC2 instance to oversee Docker hosts, enhancing the deployment workflow for containerized applications
NextConf
Developed a secure video conferencing application using Next.js, TypeScript, Clerk, getstream, shadcn, and Tailwind CSS.Implemented key functionalities such as meeting creation, recording, screen sharing, and participant management, along with a responsive design to ensure usability across devices
RentRipple
RentRipple is a marketplace platform connecting property owners and renters with backend APIs, authentication, real-time chat, MongoDB schemas, containerized services, and team-led code reviews.
GreenCart
GreenCart is an e-commerce website dedicated to selling eco-friendly and sustainable products, promoting greener choices for everyday living.
MediaChain
Its a Hyperledger Fabric based project.Its a Blockchain application to distribute royalties for the entertainment industry.
MultiClient Reverse Shell
This project is a proof of concept (POC) worm and ransomware that exploits the psexec vulnerability in Windows and SSH vulnerability in Linux systems to download and execute ransomware files, similar to WannaCry and Mirai.
A stack tuned for AI systems that need to keep running.
Also comfortable with C, C++, Java, JavaScript, SQL, GO, Bash, React, Express, Spring Boot, Flask, Django, NestJS, Langfuse, HTML, CSS, PHP, MySQL, MS SQL, MongoDB, PostgreSQL, Redis.