Anup Bhowmik

Hi, my name is

Anup BhowmikAI Enthusiast & Researcher

I am a Software Engineer with 3+ years of experience in FullStack Development and 1+ year of experience in DevOps and AI Integration in SaaS products. I have a strong foundation in practical Machine Learning and front-end development. I have experience delivering production-grade solutions across Web Apps, DevOps, and Cloud Infrastructure.

Work Experience

Software Engineer-II@Pridesys

Dhaka, Bangladesh | June 2024 - Present

  • Developed an Agentic AI system with Retrieval-Augmented Generation (RAG) using LangChain, enabling real-time data retrieval and enhancing response accuracy in a data-heavy ERP system.

  • Self-hosted local Large Language Models (LLMs) using Ollama, reducing dependency on external APIs and improving data privacy.

  • Developed an AI-driven business insight generation system, leveraging machine learning and Business Intelligence(BI) tools to provide real-time analytics, reports and decision support.

  • Set up LLM analytics and tracing with a self-hosted Langfuse instance, improving observability, performance monitoring and cost analysis.

  • Deployed the backend and database of the cloud ERP on ACS (Azure Cloud Service) using AKS (Azure Kubernetes Service), ACR(Azure Container Registry), and Cloud storage.

  • Incorporated CI/CD with GitHub actions, accelerating integration and deployment across multiple microservices.

  • Deployed the Webapp in Azure Static Web Apps. Currently leading the development of frontend (React) application and Landing page (Next.js), focusing on performance and user experience.

  • Mentoring and training a team of 4 trainee software engineers, guiding them in software development best practices and project workflows.

  • Conducted multiple training sessions on Java as part of the EDGE Project, delivering technical instructions to Trainee Developers.

LangChain
Ollama
Langfuse
React
Next.js
Azure
HashiCorp Vault
GitHub Actions
Kubernetes
Docker
PostgreSQL

DevOps Engineer@Brainlytic

Dhaka, Bangladesh | February 2025 - Present

  • Implemented CI/CD pipelines using Watchtower, Docker, and GitHub Actions, reducing deployment time by 90% and increasing system reliability.

  • Configured Traefik as a reverse proxy and automated SSL certificate management with Let's Encrypt, improving security and reducing manual intervention.

  • Conducted load testing and performance analysis using k6 and Apache JMeter. Identified bottleneck APIs and enhanced system scalability by 50% by suggesting new cloud infrastructure design.

  • Managed DNS configurations in Cloudflare, optimizing domain security, traffic routing, and reducing latency.

Docker
Traefik
watchtower
GitHub Actions
k6 Grafana
JMeter

Research Assistant & Developer@IWFM

Dhaka, Bangladesh | August 2022 - November 2022

  • Developed A web-based early warning system to warn the households living in coastal regions of Bangladesh in case of river erosion.

  • Integrated Google Maps API to visualize real-time riverbank erosion data, providing users with up-to-date geographical insights.

  • Deployed and maintained the live system, accessible at ews-re.com.

Google Maps API
React
PostgreSQL
Node.js
Firebase

My Skills

Languages

  • Python
  • Java
  • C/C++
  • JavaScript
  • TypeScript
  • HTML
  • CSS
  • Bash
  • x86 Assembly

AI/ML

  • LangChain
  • LangFuse
  • TensorFlow
  • PyTorch
  • Scikit-learn
  • OpenCV
  • Pandas
  • NumPy

DevOps and Cloud

  • Docker
  • Kubernetes
  • Minikube
  • Vault
  • Terraform
  • Microsoft Azure
  • GitHub Actions
  • Watchtower
  • Traefik
  • k6
  • Grafana
  • JMeter

Frontend

  • React
  • Next.js
  • TailwindCSS
  • ESLint
  • i18n
  • Android Studio
  • Flutter

Backend

  • Node.js
  • Spring Boot
  • FastAPI

Databases

  • PostgreSQL
  • Firebase
  • Oracle

Some Things I have Built

ews-re

Featured Project

Early Warning System River Erosion

A web-based early warning system to warn the households living in coastal regions of Bangladesh in case of river erosion. This project is under IWFM, BUET and funded by ICT Division Bangladesh.

Features

  • Integrated Google Maps API with GeoJSON to visualize real-time riverbank erosion data.
  • Developed a real-time visualization system (using Polygons) to display river erosion levels categorized as high, medium, or low.

Technical Details

  • Developed a web application using React and Node.js.
  • Generated River Erosion Hazard Map from satellite imagery and measured cross-sectional data by analyzing time series Sediment data.
  • Converted the river erosion hazard map into GeoJSON format and stored in Firebase and served via the Node.js app.
  • Google Maps API
  • React
  • PostgreSQL
  • Firebase
  • Node.js
neural-network

Featured Project

Neural Network From Scratch

Implemented a neural network using Python. The implementation includes key features such as model architecture, training, evaluation, and optimization. The project is structured to be modular and easily extensible.

Features

  • Customizable Model Architecture
  • Implemented training loops with backpropagation and optimizers
  • Hyperparameter Tuning
  • Batch size modification and optimization techniques (ADAM optimizer)

Model Definition

  • Input Layer
  • Hidden Dense Layers (with ReLU as activation function)
  • Dropout Layer (for regularization)
  • Output Layer (for classification or regression tasks)

Training Process

  • Forward propagation
  • Loss computation (cross-entropy for classification)
  • Backpropagation
  • Parameter updates using ADAM optimizer
  • EMNIST Dataset
  • Classification
  • Python
  • Machine Learning
minesweeper

Featured Project

Minesweeper AI Solver

Implemented the classic Minesweeper game in Python with an AI solver that uses logical inference to make intelligent decisions. It features a graphical interface built with PyGame and a knowledge-based AI that simulates human-like reasoning to solve the puzzle.

Features

  • Interactive PyGame-based GUI
  • AI assistant that plays using propositional logic
  • Console output of AI's reasoning steps

AI Logic

  • Knowledge base using logical sentences
  • Inference rules for safe/mine cell deduction
  • Smart move selection with fallback to random guessing
  • Python
  • AI
  • Logical Inference
  • Knowledge Engineering
  • PyGame
microcontroller-game

Featured Project

Game on Microcontroller

Implemented a simple yet engaging Space Attack game on an ATMega32 microcontroller. The game uses a button-based controller and an LED matrix display, with an efficient hardware interface powered by shift registers.

Features

  • Real-time gameplay on 8x8 LED Matrix display
  • Three-button controller (move up, move down, shoot)
  • LCD display for score and health tracking
  • Efficient pin usage via shift registers

Technical Highlights

  • Applied Daisy-chain shift registers to reduce I/O pin usage from 96 to just 3. We can send 8 bits of data serially (one by one) using just 3 control pins from the microcontroller.
  • Compatible with Microchip Studio and Proteus for simulation and hex file generation
  • Microcontroller
  • Shift Register
  • ATMega32
  • LED Matrix
dino-game

Featured Project

Chrome Dino Game

This project was developed for a workshop on PyGame fundamentals, where participants built a Chrome-style Dino Game from scratch. The game involves a dinosaur dodging obstacles with increasing speed, featuring animations, sound, and real-time score tracking.

Features

  • Chrome Dino-style endless runner
  • Jump and duck mechanics using keyboard input
  • Increasing difficulty over time
  • Real-time score tracking
  • Sound effects and background music

Workshop Concepts Covered

  • Rendering assets and animations
  • Event handling in PyGame
  • Movement and collision detection
  • Playing audio with PyGame mixer
  • Event handling
  • Audio Mixer
  • Render Animation
  • Python

Other Noteworthy Projects

Logistic Regression and AdaBoost Classifier

Implemented Logistic Regression and AdaBoost classifiers from scratch. Focused on understanding the inner workings of both models and evaluated them across multiple real-world datasets (Telco Customer Churn, Adult and Credit Card Fraud Detection).

  • Machine Learning
  • Classifier
  • Python

Assembly Code Templates

A collection of beginner-friendly Assembly language templates for practicing fundamental programming concepts such as I/O, branching, loops, and array operations. Includes mini-projects like insertion sort and binary search.

  • 8086 Assembly
  • Sorting
  • Binary Search

Software Engineering Design Patterns

A structured set of problem solutions demonstrating various software engineering design patterns with UML diagrams and unit testing. Covers OOP, Creational, Structural, and Behavioral (Observer) design patterns.

  • Software Engineering
  • OOP
  • Java

Creative Production Management

A comprehensive software engineering project designed to manage production, client interactions, budgeting, and designer performance. The system includes full lifecycle design from requirement gathering and diagramming to implementation.

  • ERD
  • BPMN
  • State Diagram

Compiler from Scratch

Built a C compiler using Flex and Bison, featuring a symbol table, lexical analyzer, syntax analyzer, semantic analyzer, and intermediate code generation with optimizations. This compiler supports a C-like language and produces 8086 assembly code.

  • Flex
  • Bison
  • Lexical Analysis
  • ICG

Basic DSA

Foundational data structures (Queue, Stack, Heap, BST, Linked List), classic sorting (Merge, Quick, Insertion), greedy techniques, and dynamic programming problems like Knapsack, Edit Distance, and Weighted Job Scheduling—ideal for learning and interviews.

  • Greedy
  • DP
  • Sorting
  • BST
  • Heap

Advanced DSA

Concise implementations of key Data Structures (Hash Table, Binomial Heap) and Graph Algorithms including traversal (BFS, DFS), Dijkstra, Bellman-Ford, MST (Prim, Kruskal), SCC, Topological Sort (Kahn’s Algo), and Ford-Fulkerson. Great for quick revision, interviews, and competitive programming.

  • DSA
  • C++
  • Graph Algo

Numerical Analysis

Implemented key numerical methods including Solving Nonlinear Equations (Newton-Raphson Method), System of Linear Equations(Gaussian Elimination), Interpolation(Newton’s Divided Difference Interpolation, Lagrangian Interpolation), Integration(Simpson’s 1/3rd Rule), and Linear and Non-linear Regression Analysis.

  • Regression
  • Interpolation
  • Python

Latin Square CSP Solver

A CSP solver for Latin Square puzzle using backtracking with domain reduction and constraint checks. Supports variable ordering heuristics (minimum remaining values, max forward degree, random) and value ordering. Efficiently enforces row-column uniqueness through scoped domain updates.

  • AI
  • Backtracking
  • Python

N-Puzzle AI Solver

Solves the classic N-Puzzle using A* Search. Uses Hamming and Manhattan heuristics for optimal pathfinding. Priority Queue and BFS are used for efficient state exploration.

  • AI
  • A* Search
  • Priority Queue
  • BFS

My Educational Background

bsc

BSc in CSE

Bangladesh University of Engineering and Technology (BUET)

April 2019 – July 2024

CGPA: 3.84/4.00

What's Next?

Get In Touch

I'm open to new opportunities in the domain of ML engineering, DevOps and cloud infrastructure! Feel free to reach out to me!