Utkarsh
Rajput
AI Researcher & Full-Stack Developer
Software-developer-turned-AI-researcher specializing in NLP, Computer Vision, and ML systems. Building AI-powered experiences with craft & curiosity.
About Me
Passionate about translating research into scalable products and mentoring budding developers
My Journey
Software-developer-turned-AI-researcher with 1-plus year of hands-on experience building production ML/DL systems and full-stack web apps. Currently pursuing a B.E. (Hons) in Software Engineering at UNSW and leading projects on cross-lingual NLP, AI-driven trading strategies and predictive analytics.
Passionate about translating research into scalable products and mentoring budding developers. I thrive at the intersection of cutting-edge AI research and practical software engineering.
Certifications
All certifications from HackerRank
Honors & Awards
3rd Place - WARTEX 7.0
Technical Competition
Top 10 - IECSE Freshers' Codesprint 2023
Programming Contest
Top 10 - IEEE-SBM WiE Codathon 2023
Coding Competition
Work Preferences
Professional Experience
Building production systems and leading teams in AI research and full-stack development
Software Engineer for AI Training (Code Quality & Debugging)
Create and answer advanced questions on software-engineering concepts to train AI models
Review AI-generated code (JavaScript, Python, Go, Java, TypeScript, C++), analysing quality, maintainability and adherence to real-world standards
Provide expert feedback on version control workflows, collaborative coding practices and debugging techniques
Contribute datasets used for full-stack development, ML and model-interpretation tasks
Software Engineering Project Lead
Re-platformed a legacy MS-Access banking stack to a secure real-time web system serving 30k+ users
Built AI-driven reporting pipelines and a corporate-spend module that cut processing time by 60%
Led full-stack development using modern web technologies and cloud infrastructure
Implemented security best practices and real-time data synchronization
AI Developer
Designed and evaluated ML models for educational technology applications
Handled feature engineering & data pipelines for large-scale datasets
Presented insights to stakeholders and contributed to product strategy
Collaborated with cross-functional teams to integrate AI solutions
Web-Developer Intern
Delivered a responsive Django-based site with modern UI/UX design
Implemented SEO & performance optimizations improving load times by 40%
Improved cross-browser compatibility and mobile responsiveness
Collaborated with design team to implement pixel-perfect interfaces
Head of Software Development
Led agile team of 12 developers across multiple AI/ML projects
Ran code reviews and established development best practices
Mentored juniors on AI/ML projects and career development
Organized technical workshops and hackathons for 200+ students
Additional Experience
Education
Academic journey spanning software engineering, AI research, and technical leadership
B.E. (Hons) Software Engineering
Advanced coursework in Machine Learning and Distributed Systems
B.E. Computer Science
AI/ML specialization with focus on Deep Learning
Published heart-disease prediction research
Head of Software Development, Enigma Technical Society
Senior Secondary Education
Physics, Chemistry, Mathematics with Computer Science
Arduino automation projects and classroom IoT systems
Academic Highlights
Leadership
Head of Software Development
Research
Published ML Research
Innovation
IoT & Arduino Projects
Featured Projects
A showcase of innovative solutions at the intersection of AI and practical applications
Local-LLM Kit
PyPI package providing an OpenAI-compatible REST/streaming API for local LLMs; supports function-calling, JSON-mode & logprobs; Dockerised with CI/CD for on-prem deployment.
Model-Compression Pipeline
End-to-end framework for pruning, quantisation, distillation & lottery-ticket search; shrinks SOTA models by ≤ 75% and doubles edge-GPU throughput.
Dynamic-NeRF
Temporal NeRF with spatio-temporal attention to render moving scenes; full training & rendering pipeline.
Continual-Learning System
Research platform implementing EWC, Experience Replay & LwF with visual forgetting metrics for lifelong learning benchmarks.
iStockz Paper-Trading Platform
MVC Python web app with yfinance feeds, secure auth and portfolio analytics.
Technical Skills
Comprehensive expertise across the full technology stack, from AI research to production systems
Languages & Frameworks
Machine Learning & AI
Data & Analytics
Web & Cloud Technologies
DevOps & Tools
Specialized Skills
Core Competencies
AI/ML Research
Deep expertise in neural networks, NLP, and computer vision with published research
Full-Stack Development
Production-ready web applications with modern frameworks and cloud deployment
System Architecture
Scalable system design, DevOps practices, and team leadership experience
Research & Publications
Cutting-edge research at the intersection of AI, NLP, and practical applications
Cross-Lingual & Temporal Dynamics in Fake News Detection
Built multilingual transformer pipeline over 44k articles (Hindi + English) with 92% F1 score
Implemented temporal-aware sampling and topic-drift robustness tests
Advanced cross-lingual transfer learning techniques
Novel approach to handling temporal dynamics in fake news detection
Neural-Network–Based Trading Strategies in Pine Script
Designing real-time NIFTY/Bank NIFTY options strategy using engineered features
CNN + LSTM hybrid architecture for market prediction
Live-data ingestion via WebSocket with real-time processing
Arrow annotations locked post-placement for strategy validation
Feature Engineering Optimisation & Stacked Weighted Model for Heart-Disease Prediction
Employed MICE imputation and Truncated SVD (9 components) for data preprocessing
Random Forest (200 trees) and ExtraTrees with advanced stacking techniques
Achieved 89% ROC-AUC on Kaggle Heart Disease dataset
Comprehensive feature engineering optimization pipeline
Research Impact
3 Publications
Active research in AI, NLP, and ML applications
92% F1 Score
State-of-the-art performance in fake news detection
Real-time Systems
Live trading algorithms and real-time data processing
Let's Connect
Ready to collaborate on something amazing? I'd love to hear about your project or just have a chat about AI, technology, and innovation.
Get in Touch
Whether you're looking to discuss a potential collaboration, have questions about my research, or just want to connect with a fellow AI enthusiast, I'm always open to meaningful conversations.
kumarutkarsh815@gmail.com
Phone
Available upon request
Location
Sydney, Australia
Response time: Usually within 24 hours