Skip to content
Portrait of Sohel AhmedSohel AhmedBackend and IoT systems, Bangladesh

I build systems that connect software, devices, and the real world.

Software developer focused on real-time systems, GPS tracking, IoT, backend infrastructure, and scalable web applications. I work from the tracker on the vehicle to the dashboard on the screen.

View projectsContact me
Tracker4GNginx TCPNode.jsRedis + MongoDBDashboardN23.7°N 90.4°E
Telemetry, tracker to dashboardCommands back to the device

I like knowing how the whole thing works.

I’m a software developer from Bangladesh. Most of what I build has a physical device at one end and a person at the other, and a lot happens in between: radio links, TCP sockets, binary protocols, queues, caches, databases, and finally a dashboard.

I don’t stay in one layer. When a tracker stops reporting, the cause can be the firmware, the 4G link, the proxy, the parser, or the cache, and I want to be able to follow it through all of them. That’s why I work across frontend, backend, infrastructure, networking, and hardware. The hard problems are usually where those layers meet, and those are the ones I find most interesting.

Portrait of Sohel Ahmed
Sohel Ahmed, Bangladesh

The layers I work across

  1. interface
  2. realtime
  3. data
  4. backend
  5. protocol
  6. network
  7. hardware

Systems with moving parts.

The kind of software where devices, networks, and people are all part of the same loop.

Real-Time Systems
Systems where data moves continuously between devices, servers, and users.
Socket.IO, WebSocket, Redis, Firebase RTDB
GPS & Fleet Platforms
Vehicle tracking, device communication, live maps, alerts, reporting, and fleet management.
JT808, GT06, Leaflet, MapTiler
Distributed Backend Systems
Node.js services, Redis, MongoDB clusters, queues, workers, and scalable infrastructure.
Node.js, Replica sets, Sharding, Workers
IoT Systems
Connecting sensors, microcontrollers, GPS devices, BMS systems, and backend services.
ESP32, MicroPython, SIM7600G-H, BMS
Infrastructure
Docker Swarm, Nginx, Linux VPS, networking, reverse proxies, TCP load balancing, and security.
Docker Swarm, Nginx, Tailscale, Fail2Ban
Protocol Engineering
JT808, GT06, TCP, UDP, WebSocket, UART, RS485, and CAN.
TCP, UDP, UART, RS485, CAN

From a tracker on a vehicle to a fleet manager’s screen.

A simplified view of the GPS platforms I build and run. Select a stage to see what happens there.

  1. Stage 1 of 9, device layer

    GPS Tracker

    The vehicle-side device. It reports position, status, and alarms, and accepts commands back from the platform.

    • Position uploads
    • Heartbeats
    • Alarm flags
    • Command responses

Production systems, protocols, and hardware.

What each system does, and what made it hard to build.

GPS Tracking · Fleet Management · IoT Platform

Sultan Tracker

A GPS tracking and fleet management platform in Bangladesh. It covers the full chain: tracker devices talking TCP, live locations, alerts, fleet management, and real-time dashboards for operators.

The hard part

Keeping thousands of concurrent GPS tracker connections alive and parsed in real time. Behind Nginx TCP proxying and a distributed deployment, every incoming position has to be decoded, cached as live state, persisted, and pushed to the dashboards that are watching it.

What it does

  • TCP server infrastructure
  • GPS data parsing
  • Real-time location processing
  • Redis-based live location caching
  • MongoDB storage
  • Alert processing
  • Device command processing
  • Real-time dashboards
  • Distributed deployment
  • Nginx TCP proxying & load balancing

Built with

Next.js, React, TypeScript, Node.js, Express.js, MongoDB, Redis, Firebase RTDB, Socket.IO, Docker Swarm, Nginx, Leaflet, MapTiler

IoT · GPS · Protocol Engineering

JT808 Tracker Platform

A tracker backend built on the JT808 protocol. It receives device data over TCP, parses JT808 packets, processes positions, calculates mileage, stores device state, and talks back to the trackers.

  1. 7Eflag
  2. 02 00msg id
  3. ·· ··props
  4. ·· ·· ··terminal
  5. ·· ··serial
  6. body…position
  7. ··xor
  8. 7Eflag
Message 0x0200, a location report, escaped and checksummed

The hard part

Implementing a binary protocol correctly at the byte level. That includes framing, escaping, message IDs, and serials, plus keeping each device's session and state consistent when the connection drops and reconnects.

  • JT808 packet parsing
  • Device session management
  • GPS position processing & mileage
  • Command queue & terminal parameter config
  • Persistent device state across reconnects

Node.js, TCP, JT808, MongoDB, Redis, Firebase, Docker

Embedded · IoT · Hardware

Battery & IoT Prototyping

Ongoing hardware work on lithium battery systems: BMS integration, GPS-enabled battery monitoring, temperature sensing, and remote charge and discharge control from a backend.

  1. Cells
  2. BMS
  3. ESP32
  4. Backend
BMS and ESP32 talk over UART, RS485 or CAN: telemetry up, charge control down

The hard part

Making physical systems such as cells, BMS boards, relays, and sensors speak reliably to software over UART, RS485, and CAN, then making them controllable remotely.

  • BMS communication
  • GPS-enabled battery monitoring
  • Remote charge / discharge control
  • Temperature monitoring
  • Relays & sensors

ESP32, MicroPython, BQ76920, JK BMS, Daly BMS, GPS modules, RS485, CAN, UART

Real-time Communication · CRM

WhatsApp CRM

A WhatsApp-based CRM and backend for handling customer conversations. Messages, status updates, and CRM workflows stay in sync across agents in real time.

The hard part

Bridging an external, webhook-driven messaging API with a live agent UI, so that conversations and message status stay consistent for everyone looking at them.

  • Customer conversations
  • Real-time messaging
  • Message status tracking
  • CRM workflows
  • Backend API

Next.js, Express.js, MongoDB, Redis, Socket.IO, WhatsApp Cloud API

Fleet Tracking · Reporting

Rangs Tracker

A tracking and reporting application for vehicle data. It provides filterable dashboards and monthly reports built from GPS tracking records.

The hard part

Turning large volumes of raw tracking records into reports and views that are fast to filter and easy to read.

  • Monthly reports
  • Filtering
  • Dashboard UI
  • Vehicle data visualization

Next.js, React, TypeScript, MongoDB

Tools I reach for first.

The core set I use day to day, followed by the wider set I work with across the stack.

  • Next.js
  • React
  • TypeScript
  • Node.js
  • Express
  • MongoDB
  • Redis
  • Docker
  • Nginx
  • Linux
  • Firebase
  • Socket.IO
  • 0x7EJT808
  • ESP32
Frontend
JavaScript, Tailwind CSS, shadcn/ui, Zustand, App Router, Leaflet, MapTiler, Real-time dashboards, Responsive UI
Backend
REST APIs, WebSocket, JWT auth, Background workers, Command processing, TCP servers
Database
MongoDB replica sets, MongoDB sharding, Firebase Realtime Database
Infrastructure
Docker Compose, Docker Swarm, Ubuntu, DigitalOcean, Contabo VPS, Tailscale, Certbot, DNS, Reverse proxy, TCP proxying, Load balancing, Monitoring, Fail2Ban, Asterisk / PJSIP, TURN / coturn
IoT / Embedded
MicroPython, Raspberry Pi, SIM7600G-H 4G, GPS trackers & modules, BMS, TPMS, Relays, Sensors, PCA9685, MQTT concepts, Hardware debugging
Protocols
GT06, TCP, UDP, HTTP, UART, RS485, CAN, RTP / SIP
Android
Native Android, Java, GPS / tracking apps, Device communication

How I approach the work.

  • Understand the system, not just the API.

  • Build for real-world constraints.

  • Measure before optimizing.

  • Keep infrastructure observable.

  • Hardware and software are one system when you’re building IoT.

Each layer led to the one below it.

I started with interfaces. Each project pushed me a layer further down, toward the device.

  1. Web Development

    React, Next.js, and TypeScript. Interfaces, state, and the browser as a platform.

  2. Backend Systems

    Node.js and Express APIs, authentication, MongoDB, and the work of keeping data correct.

  3. GPS Tracking

    Tracker devices, TCP servers, and binary protocols like JT808 and GT06. This is where the work stopped being only HTTP.

  4. Real-Time Infrastructure

    Redis, Socket.IO, and Firebase. Moving live state from devices to dashboards as it happens.

  5. IoT / Embedded Systems

    ESP32, MicroPython, BMS boards, UART, RS485, and CAN. Working on the physical side of the system too.

  6. Distributed Systems

    Docker Swarm, Nginx TCP load balancing, and sharded MongoDB. Making the whole chain hold up under load.

Have a system that needs to be built? Let’s talk.

Email is the fastest way to reach me. Tell me what the system has to do and where it runs.

sohel.ahmed2178@gmail.com