← Back to list
Registration: 02.07.2026

Tushar Debbarma

Specialization: Applied AI Engineer
Systems-minded Applied AI Engineer with 1.8+ years of experience in mission-critical industrial automation (PLC/SCADA/DCS) and deep learning applications. — Expert in bridging OT/IT gaps through real-time optimization and computer vision systems. — Delivered ~19% reduction in plant power consumption via VFD recalibration using Taguchi statistical methods at HPCL. — Fine-tuned CNN via transfer learning for Alzheimer's Disease detection, cutting false positives by 85%. — Proficient in Python, TensorFlow, Keras, OpenCV, and Scikit-learn. — B.Tech in Electronics & Instrumentation Engineering from National Institute of Technology, Agartala.
Python SQL Git Machine Learning JavaScript flask Linux pandas AWS C++ Node.js Matplotlib MongoDB OpenCV scikit-learn Docker MySQL Tensorflow Java React
Systems-minded Applied AI Engineer with 1.8+ years of experience in mission-critical industrial automation (PLC/SCADA/DCS) and deep learning applications. — Expert in bridging OT/IT gaps through real-time optimization and computer vision systems. — Delivered ~19% reduction in plant power consumption via VFD recalibration using Taguchi statistical methods at HPCL. — Fine-tuned CNN via transfer learning for Alzheimer's Disease detection, cutting false positives by 85%. — Proficient in Python, TensorFlow, Keras, OpenCV, and Scikit-learn. — B.Tech in Electronics & Instrumentation Engineering from National Institute of Technology, Agartala.
Python SQL Git Machine Learning JavaScript flask Linux pandas AWS C++ Node.js Matplotlib MongoDB OpenCV scikit-learn Docker MySQL Tensorflow Java React

Portfolio

Whitener-Detection-Model

Project deploys YOLOv3-based AI to detect correction pen fluid in images via a user-friendly Flask app, aiding precision in various industries.

Reinforcement Learning — Frozen Lake Environment

Built and trained a Reinforcement Learning model (Q-Learning) to solve the Frozen Lake environment. Implemented the agent from scratch using Python and NumPy, optimizing the reward policy through iterative training with epsilon-greedy exploration. Analyzed convergence behavior across multiple episode runs.

AI upscaling model

A Python-based real-time video streaming engine capable of ingesting multiple video sources, upscaling them to 4K, transcoding using Apple Silicon hardware acceleration, generating HLS output, and serving the stream through a local web server. The long-term goal of this project is to build an AI-powered streaming platform with live subtitle translation, AI video enhancement, and low-latency streaming.

Skills

Python
TensorFlow
Keras
OpenCV
Scikit-learn
NumPy
Pandas
Matplotlib
Machine Learning
Deep Learning
Computer Vision
PLC
SCADA
DCS
C/C++
Java
Docker
AWS
Git
MySQL
MongoDB
Flask
Linux
SQL

Work experience

Terminal Automation Officer
08.2024 - 03.2026 |Hindustan Petroleum Corporation
PLC, SCADA, DCS, BCU, MOXA Device Servers, VFD, Python, TensorFlow, Taguchi Methods
● Architected and maintained critical Terminal Automation Systems (PLC/SCADA/DCS), ensuring zero-downtime system reliability and data-driven process optimization. ● Delivered ~19% reduction in plant power consumption by recalibrating VFDs using Taguchi statistical methods. ● Executed automation system upgrades and retrofits on live, high-stakes infrastructure without any downtime. ● Identified and eliminated bottlenecks at underperforming loading points, translating operational data into actionable efficiency improvements.
CNS Systems Engineer
2023 - 2023 |Airports Authority of India
CNS Systems, Radio-frequency Systems, Electromagnetic Spectrum Management
● Supported fail-safe mechanisms for Communication, Navigation & Surveillance (CNS) systems to reduce data-transmission risk. ● Studied radio-frequency and electromagnetic-spectrum management to prevent interference and channel congestion in air traffic communications.
Undergraduate Researcher
2023 - 2023 |National Institute of Technology, Agartala
Python, TensorFlow, Keras, Matplotlib, Seaborn, CNN, Transfer Learning
● Fine-tuned a CNN via transfer learning for Alzheimer's Disease detection, reducing false positives by 85%. ● Built deep learning pipeline in Python with TensorFlow and Keras, using Matplotlib and Seaborn for results visualization and analysis.

Educational background

Electronics and Instrumentation Engineering (Bachelor’s Degree)
2020 - 2024
National Institute of Technology, Agartala

Languages

HindiAdvancedBengaliAdvancedEnglishAdvanced