$ whoami

Ayush Mali

Fullstack Software Engineer

Building scalable distributed systems, event-driven architectures, and intuitive user interfaces with Go, Python, and TypeScript. // Write once. Deploy anywhere. Sleep peacefully.

A quick audio summary about me

$ cat about.txt

╭─ About Ayush Mali

Fullstack Software Engineer with expertise in building concurrent systems, event-driven architectures, and modern web applications. I specialize in Go for high-performance backend services, Python for data processing and ML, and TypeScript/React for intuitive frontends.

My experience spans from low-level binary protocol parsing and PLC integrations to cloud-native Kubernetes deployments with GitOps workflows. I'm passionate about infrastructure automation, test-driven development, and building systems that are both performant and maintainable.

"Any fool can write code that a computer can understand. Good programmers write code that humans can understand." - Martin Fowler

╰─ EOF

$ ls -la skills/

Languages

→ Go
→ Python
→ TypeScript
→ JavaScript
→ SQL
→ Bash

Backend & Systems

→ FastAPI
→ Flask
→ Gorilla Mux
→ Express
→ Linux/UNIX

Frontend

→ React
→ Next.js
→ Storybook
→ Vitest
→ Cypress

Cloud & DevOps

→ Docker
→ Kubernetes
→ Ansible
→ Argo CD
→ Azure
→ Prometheus
→ Loki
→ Grafana
→ Proxmox
→ KVM
→ QEMU

Databases & Messaging

→ PostgreSQL
→ SQLite
→ MongoDB
→ Redis
→ Kafka
→ MQTT

Machine Learning

→ scikit-learn
→ Pandas
→ NumPy
→ NLTK
→ Random Forest

$ find ./projects -type f -name "*.md"

Nimbus Store

A production-grade, event-driven e-commerce platform built with a microservices architecture, featuring personalized product recommendations, real-time data processing, and secure payment workflows.

Problem:

Design and build a scalable e-commerce system with ML-powered recommendations, real-time event processing, secure authentication, and encrypted data flows.

Tech Stack:

FastAPINext.js 14MongoDBKafkaRedisGraphQLscikit-learnNLTKPandas

Code Demo

Reddit Scraper

An open-source Reddit scraping tool that collects subreddit, post, comment, reply, and user data using only Python standard libraries, designed for reliability and large-scale data ingestion.

Problem:

Efficiently scrape Reddit data at scale while handling API rate limits using a custom exponential backoff strategy.

Tech Stack:

PythonBashMongoDB

Code

DS Playground

A curated collection of data structures and algorithms with brute-force, optimized, and optimal solutions, accompanied by an interactive web interface.

Problem:

Systematically solve and document popular DSA and LeetCode-style problems with clear progression from naive to optimal approaches.

Tech Stack:

PythonBashNext.js

Code Demo

xhibit

A lightweight CLI tool that displays system information alongside ASCII art or custom images directly in the terminal.

Problem:

Provide a customizable and visually appealing way to display system diagnostics on Linux-based operating systems.

Tech Stack:

PythonBash

Code

sYT

A terminal-based YouTube video browser and player that enables video selection and playback without relying on the official YouTube API.

Problem:

Enable video discovery and playback from the terminal using Unix tools while avoiding dependency on YouTube APIs.

Tech Stack:

Bashfzfueberzug

Code

Screenlocker

A screensaver-style application that displays Apple Aerial-style videos as a dynamic lock screen experience.

Problem:

Create a visually rich screensaver using video processing and native performance for desktop environments.

Tech Stack:

RustOpenCV

Code

ghkd

A system-level hotkey daemon for Linux that captures keyboard input directly from evdev, enabling global keybindings across Wayland, X11, and TTY environments without compositor dependencies.

Problem:

Build a display-server–agnostic hotkey system capable of handling global keyboard shortcuts reliably at the kernel input layer while maintaining minimal overhead and safe device detection.

Tech Stack:

GoLinux evdevYAMLSystem Programming

Code

autost

A declarative Linux session autostart manager that executes startup commands and inline scripts from a validated YAML configuration for reproducible desktop environments.

Problem:

Replace fragile shell-based startup scripts with a structured, validated system that manages environment variables, commands, and scripts in a predictable and maintainable way.

Tech Stack:

GoYAMLLinuxProcess ManagementXDG Specification

Code

$ history | grep experience

Fullstack Software Engineer

Midaas Telesoft Pvt Ltd • Sept 2024 - Present

→Concurrent SMS Processing: Designed a concurrent Go system to process SMS from multiple modems in parallel, extract OTPs, and persist results in SQLite. Built a Next.js dashboard to visualize real-time modem health, signal quality, and network state.
→VPN Management System: Architected VPN lifecycle management for OpenVPN and WireGuard with CRUD operations, auto-reconciliation, and state synchronization via Linux NetworkManager D-Bus APIs. Implemented ovpn file parser, secure key and certificate storage, and a UI for VPN configuration and status monitoring.
→Cloud Cover Prediction: Trained a Random Forest classifier on sky image frames, achieving ~80% accuracy for cloud coverage prediction. Deployed a Flask API returning cloud percentages and Okta values for meteorological analysis.
→TwinCAT PLC Integration: Implemented Go parsers to decode all Beckhoff TwinCAT PLC datatypes from raw binary data and enumerate available symbols. Used Go's encoding/binary package for primitive types and custom parsers for complex structures.
→Apparatus Monitoring System: Built a thread-safe MQTT ingestion system in Go using Eclipse Paho and JetStream to monitor circuit breaker and isolator signals from 16-channel DI modules. Integrated Pioneer Wincon wind turbine telemetry capturing HVRT, LVRT, wind speed, and wind direction with event-driven SQLite logging and a Next.js configuration UI.
→Reusable UI Components: Built single and multi select combobox components in React supporting static, generic, and async data sources. Followed test-driven-development using Vitest and documented the components with Storybook.
→Infrastructure Automation: Automated MikroTik switch configuration using idempotent Ansible playbooks covering VLANs, ports, MLAG/LACP bonds, routing, and DNS. Provisioned multi-node k0s Kubernetes clusters with Cilium CNI and Longhorn CSI, and managed deployments via Argo CD GitOps.

$ ./contact.sh

Get In Touch

Interested in discussing algorithms, system design, or potential collaborations? Feel free to reach out through any of the channels below.

ayushm360o@gmail.com github.com/glowfi

Send Message