Ayush Sinha

Full Stack Engineer with 3+ years building production systems. I work at Yotta Data Services on Apiculus, a multi-tenant cloud management platform, across the Vue frontend, the Node.js services behind it, and the Apache CloudStack engine underneath. Most of my recent work has been performance debugging: finding why something is slow, or why it only breaks under load.

šŸ’¼ MY JOURNEY

Upstream Contributions

Open Source

Apache CloudStack
āœ“Merged

Apache Software Foundation Ā· Release 4.22.2 Ā· PR #13700 Ā· Sep 2026

Fixed a concurrency race that silently left VM network interfaces attached

The Bug

CloudStack is the orchestration engine underneath Apiculus. If two API calls detached different NICs from the same VM at the same time, the async work-job queue treated them as duplicates and ran only one of them. Both calls still returned success, so the VM kept a NIC that the control plane had already written off as removed.

The Fix

The pending-work-job lookup keyed only on VM ID, the same mistake that had been fixed on the attach path years earlier. I re-keyed the deduplication on the NIC UUID so two detaches of different NICs stop colliding, and added regression tests that fail on the old code and pass on the new. One of the CloudStack committers reproduced the race on a live deployment before approving: two detach calls five seconds apart, both returning success with the NICs still attached, both cleared once the patch was in.

3

Regression tests

3

Committer approvals

149

CI smoke tests green

+106/-3

Lines changed

JavaApache CloudStackConcurrencyAsync Job Queue
View PR #13700

What I've Built

Selected Projects

github.com/Andr0human/Hammr
Hammr

01

PROJECT 01

Hammr

Load-testing tool that finds where your API breaks, not just how fast it is.

1000+

Concurrent VUs

63%

Scaling efficiency detected

3.8x

p95 latency regression caught

10

Auto-diagnostic rules

Built a load-testing tool that sustains 1000+ concurrent virtual users per process by running async VU loops on Node.js Worker Threads with pooled undici agents, rather than a thread per user. I pointed it at a production Vercel/MongoDB-Atlas API and found a capacity knee between 50 and 100 VUs: doubling load from 100 to 200 VUs bought only 1.26x more throughput (63% scaling efficiency), while p95 latency grew 3.8x, from 1.6s to 6.0s, and errors hit 8%. A rules engine sits on top with 10 rules, some per-run and some across runs, so those knees and latency regressions get flagged from the raw metrics instead of eyeballed off a graph.

TypeScriptNode.jsundiciExpressSocket.IOReactNext.jsClickHouseDockerAWS

02

PROJECT 02

Chessmate

Real-time multiplayer chess powered by Elsa, a self-built 2500+ Elo C++ engine.

50+

Concurrent rooms

2500+

Elo (Elsa engine)

95%

Lighthouse score

C++

Custom engine

Built a real-time multiplayer chess platform on WebSockets (Socket.IO) that holds 50+ concurrent game rooms, with a frontend scoring 95% on Lighthouse across devices. Single-player mode plays against Elsa, a C++ engine I wrote myself. Its v3.0.0 release measures around 2516 CCRL-blitz-equivalent Elo, up 134 Elo on the previous version over 2505 head-to-head games, and the difficulty is adjustable.

JavaScriptReactNext.jsSocket.IOChess.jsTailwind CSS
https://chessmate.ayushsinha.dev
Chessmate
https://cptrackr.ayushsinha.dev
CPTrackr

03

PROJECT 03

CPTrackr

One dashboard for every competitive programming contest across the major judges.

3

Platforms unified

Live

Contest feed

TS

Typed end-to-end

REST

Normalized API

Built a unified contest aggregator that normalizes live competition data from Codeforces, CodeChef, and LeetCode into a single dashboard. Designed a Node.js/Express backend in TypeScript to handle inconsistent third-party APIs, with a React frontend for filtering contests by platform, date, and duration.

JavaScriptReact.jsNode.jsExpressTypeScript

04

PROJECT 04

JobNest

Full-stack job tracker with OAuth and analytics across application stages.

OAuth

Secure auth flow

REST

Modular API

MongoDB

Flexible schema

Analytics

Stage funnel

Built a full-stack job application tracker with OAuth-based authentication and an analytics dashboard for tracking application status across stages. Designed a modular RESTful API in Express with MongoDB, and implemented dynamic filtering and status workflows on the React frontend.

React.jsNode.jsExpressAnt DesignMongoDB
https://jobnest.ayushsinha.dev
JobNest
github.com/Andr0human/SignStream
SignStream

05

PROJECT 05

SignStream

Real-time sign language recognition from webcam input using a CNN-LSTM pipeline.

CNN-LSTM

Model architecture

Realtime

Webcam inference

MediaPipe

Landmark extraction

TF

Custom trained model

Trained a CNN-LSTM model for real-time sign language recognition, using MediaPipe to extract hand landmarks from live video frames and feeding them into a custom TensorFlow model for gesture classification. Optimized the inference pipeline to achieve low-latency predictions on live webcam input.

PythonTensorFlowOpenCVMediaPipe

My Skills

Technologies I use across production systems, side projects, and beyond.

Production

Used at Yotta & Opslyft
AWSApache CloudStackVue.jsJavaNode.jsReactTypeScriptClickHouseRabbitMQMySQLKubernetesGrafanaOkta

Core Stack

JavaScriptTypeScriptReactNext.jsNode.jsExpressNestJS

Databases

MongoDBPostgreSQLMySQLRedisClickHouse

DevOps & Tools

DockerKubernetesGitGitHub ActionsAWSGrafanaCloudWatch

Familiar

PythonDjangoGraphQL

CONNECT WITH ME!

CodeChef
LeetCode
Hackerrank
Ayush Portfolio

Copyright Ā© 2026 Ayush - All Rights Reserved.