// Systems & AI Engineer
PULKIT
SARDANA
Creator of Solvix: Engineering high-precision computer vision models for automated solar panel defect recognition.
pulkitsardana@os:~
➜
C++
Python
FastAPI
ONNX
AVAILABLE FOR FREELANCE WORK
[CLICK ANYWHERE TO INITIALIZE AUDIO SUBSYSTEM]
PULKIT
//DEPLOYED_ARCHITECTURES
AD²-Restormer
Deep Learning / Computer Vision
A Transformer-based image restoration pipeline for rain, fog, snow, and haze degraded images.
Result: Built a modular PyTorch training and evaluation workflow for multi-weather restoration experiments.
PythonPyTorchTransformersOpenCVAWS EC2S3
Solvix
[SOURCE]Computer Vision / Research
A solar panel defect recognition system using real annotated images and interpretable visual features.
Result: 89.82% crop-level accuracy and 92.57% defect-only accuracy on a 6,493-image Roboflow dataset.
PythonOpenCVScikit-learnNumPyRoboflow
Intelligent Recommendation Engine
Machine Learning
A recommendation workflow combining content-based similarity and collaborative filtering ideas.
Result: Built an end-to-end ML pipeline for feature engineering, similarity scoring, ranking, and evaluation.
PythonPandasNumPyScikit-learn
WiseMart
[SOURCE]Full Stack + Recommendations
A React and Flask e-commerce project with authentication, cart flow, checkout, and product recommendations.
Result: Connected a practical shopping interface with recommendation models for similar and frequently purchased products.
ReactFlaskPythonJavaScriptJWT
Portfolio Website
Frontend / Personal Portfolio
A polished personal portfolio website built with Next.js, Tailwind CSS, and animated sections to showcase projects and experience.
Result: Created a modern, responsive portfolio experience with a modular component structure and smooth motion-driven content presentation.
Next.jsReactJavaScriptTailwind CSSFramer Motion
CurviMamba
[SOURCE]Deep Learning / Medical Vision
A Mamba-SSM based deep learning framework for detecting long, flexible, and deformable curvilinear structures in underwater imagery.
Result: Built a research-oriented pipeline with custom topology-aware loss, Hessian-based attention, and reproducible evaluation against baseline methods.
PythonPyTorchMambaComputer VisionDeep Learning
SentinelAI
[SOURCE]AI / Full Stack / MLOps
An enterprise-grade fraud intelligence platform that streams transactions through an AI pipeline and surfaces explainable risk insights in a live dashboard.
Result: Built a production-inspired end-to-end system with real-time ingestion, queue-based processing, explainability, and a polished monitoring interface.
Next.jsNode.jsExpressTypeScriptFastAPIPythonPostgreSQLRedisBullMQSHAPXGBoost
//SYSTEM_CAPABILITIES
AI / ML
- Machine Learning
- Deep Learning
- Computer Vision
- PyTorch
- OpenCV
- YOLOv8
- XGBoost
- Scikit-learn
- NumPy
- Pandas
- Transformers
- Model Evaluation
- SHAP
Programming Languages
- Python
- C++
- C
- SQL
- JavaScript
- Swift
Frontend & Mobile
- React
- Next.js
- SwiftUI
- HTML/CSS
Backend, APIs & Data
- Flask
- FastAPI
- Node.js
- Express.js
- REST APIs
- MySQL
- PostgreSQL
- Redis
- BullMQ
- Prisma
- Docker Compose
Cloud & DevOps
- AWS EC2
- AWS S3
- Git
- GitHub
- Cloud Computing
Core CS Fundamentals
- Object-Oriented Programming
- DBMS
- Operating Systems
- Computer Networks
Research & Domain Skills
- Image Restoration
- Image Segmentation
- Feature Extraction
- Object Detection
- Medical Imaging
- Research Experimentation
- Ablation Studies
- Paper Reproduction
- Technical Writing
- Vision Mamba
- Hessian Attention
- Topology-Aware Feature Learning
//EXPERIENCE
Data Science Trainee
July 2026 – August 2026Triple One Solutions (Course Yard Solutions Pvt. Ltd.)
Training period: 05/07/2026 – 05/08/2026
Completed a structured Data Science training program, demonstrating strong performance and contributing to technical tasks and projects during the training period.
//CERTIFICATIONS
Machine Learning Specialization
March 2026Stanford Online and DeepLearning.AI via Coursera
Completed three courses covering supervised learning, advanced learning algorithms, unsupervised learning, recommender systems, and reinforcement learning.
AI Fundamentals: Foundations for Understanding AI
July 11, 2026IBM SkillsBuild
Completed IBM SkillsBuild's AI Fundamentals: Foundations for Understanding AI learning plan. Gained foundational knowledge of AI concepts, machine learning, deep learning, neural networks, generative AI, predictive AI, computer vision, AI ethics, bias mitigation, and prompt engineering through hands-on modules and assessments.
Deloitte Technology Job Simulation
June 27, 2026Deloitte (Issued via Forage)
Completed Deloitte's Technology Job Simulation on Forage, gaining practical experience through real-world software development and coding tasks. Strengthened problem-solving, software development, and technical implementation skills by completing industry-inspired challenges that simulate the work of Deloitte technology professionals.
Web Development + DevOps + Blockchain Cohort
September 9, 2024100x Devs (Cohort 3)
Demonstrated comprehensive knowledge of Web Application Development, System Deployment, and Blockchain Based Technologies.
Google AI Certificate
July 15, 2026Completed the Google AI Professional Certificate on Coursera, developing practical skills in generative AI, prompt engineering, and AI application development. Built a portfolio of 20+ AI artifacts while learning to design effective prompts, evaluate AI-generated outputs, and apply responsible AI practices.
IBM Data Science Professional Certificate
September 19, 2026IBM Skills Network via Coursera
Completed a 12-course professional certificate covering data science methodology, Python, SQL, data analysis, data visualization, machine learning, generative AI, and an applied data science capstone project through hands-on labs and assignments.
//CONTACT
[AVAILABLE_FOR_COLLABORATION]
Let's build something useful.
Whether it's a research project, AI product idea, or a custom software build, I'm open to meaningful collaborations.
Email: sardanapulkit@gmail.com
GitHub: https://github.com/PulkitSardana