Ashish Gupta

Software Engineer

Chandigarh, India0 mo experience

Key Highlights

  • Global finalist at ICIP Conference for retinal image classification
  • First prize winner at IEEE SPS VIP Cup
  • Strong academic performance with 8.04 CGPA
Stackforce AI infers this person is a Machine Learning and Software Development specialist in the Healthcare sector.

Contact

Skills

Core Skills

Computer NetworksMachine Learning

Other Skills

Internet Protocol Suite (TCP/IP)Python (Programming Language)TensorFlowDeep LearningImage ClassificationComputer VisionWell CompletionSoftware ProjectsCooperativeImplementation PlansFront-End DevelopmentComputer EngineeringHTMLTeamworkProblem Solving

About

I’m a Computer Science Engineering undergraduate at IIT Ropar with a passion for machine learning, software development, and research-driven solutions. Over the past few years, I’ve combined academic rigor (8.04 CGPA till 6th semester) with hands-on experience to tackle real-world challenges—from medical imaging to predictive analytics and full-stack web applications. During my ML internship under Dr. Puneet Goyal for the IEEE SPS VIP Cup, I led deep learning efforts on retinal OCT image classification—denoising with Self2Self and cGANs, super-resolution via ESRGAN variants, and classification using ResNet-50 with CBAM—culminating in a global finalist presentation at ICIP Abu Dhabi and a first-place award. Currently, as a Research Intern (SDE role) at Samsung Research Institute Delhi, I’m building modular C/C++ Linux tools for per-process network monitoring, optimizing IP-to-PID mapping and bandwidth tracking inspired by nethogs. My project portfolio spans: Marker-Based Image Segmentation (DSU, Sobel operators, Flask full-stack) achieving 95% accuracy. Cognigrade: an OCR- and AI-driven answer-sheet grader with intuitive dashboards, Kafka-based notifications, and 95% grading accuracy. Dream11 Next-Gen Team Builder: an ensemble pipeline (XGBoost, CatBoost, Random Forest, KNN) with knowledge graph integration to forecast fantasy cricket performance. Interactive Cops-and-Robbers game on graph topologies (C++, SFML), and systems-level work like cache-optimization with ChampSim and a RISC-V assembler. I’m proficient in C/C++, Python, JavaScript (Flask, FastAPI), TensorFlow/PyTorch, Docker, Linux, and competitive programming (Codeforces 1428, CodeChef 1846). Recognitions include top 1% in JEE-Advanced, Institute Merit Scholarship, high CodeChef ranks, and inter-IIT challenge placements. Eager to collaborate on innovative ML/AI, software engineering, or research projects, I thrive in environments where I can learn continuously, contribute end-to-end solutions, and drive measurable impact. Let’s connect to explore opportunities and share ideas!

Experience

0 mo
Total Experience
--
Average Tenure
--
Current Experience

Samsung india

Software Engineer Intern

May 2025Present · 1 yr · Noida, Uttar Pradesh, India · On-site

  • As a Research Intern (SDE) at Samsung Research Institute Delhi, I have designed and implemented core Linux-based C/C++ modules for detailed per-process network monitoring. Drawing inspiration from the raboof/nethogs project, I developed efficient IP-to-PID mapping and bandwidth-tracking components, then refactored them into modular, standalone libraries for easy integration and reuse across multiple projects. Through careful optimization and clean API design, these tools provide accurate, low-overhead visibility into process-level network usage, facilitating diagnostics and performance analysis in diverse networking scenarios.
Computer NetworksInternet Protocol Suite (TCP/IP)

Indian institute of technology, ropar

ML Internship

Apr 2024Oct 2024 · 6 mos · Rupnagar, Punjab, India · On-site

  • ML Internship - IEEE SPS VIP Cup: Retinal OCT Image Classification:
  • Incorporated deep learning methods to classify noisy retinal OCT images into Healthy, DME, and Non-Diabetic categories.
  • Selected as one of three global finalists to present at the ICIP Conference in Abu Dhabi, where we won 1st prize.
  • Techniques adopted
  • Denoising task : Employed a Self-Supervised Self2Self Network (encoder-decoder architecture) and Conditional GAN (cGAN) to enhance image quality by reducing noise up to 90%.
  • Super Resolution: Engineered 2Encoder-1Decoder ESRGAN and Conditional GAN (cGAN) models to improve the clarity of low-resolution images in the test dataset, enhancing the quality and detail of retinal OCT images.
  • Classification: Leveraged pre-trained ResNet-50 combined with the Channel Block Attention Module (CBAM) for accurate volume classification post-denoising and super-resolution.
  • Models implemented: Conditional GAN (cGAN), Self2Self Network, 2Encoder-1Decoder ESRGAN, Through Plane SR cGAN, ResNet-50 with CBAM.
Python (Programming Language)TensorFlowMachine Learning

Education

Indian Institute of Technology, Ropar

Bachelor of Technology - BTech — Mathematics and Computer Science

Oct 2022May 2026

Central Academy Schools

High School

M. R. S Public School

Senior Secondary Education — Mathematics

Stackforce found 100+ more professionals with Computer Networks & Machine Learning

Explore similar profiles based on matching skills and experience