Aryan Gupta. Tactile Depth. Luminous Intelligence.
AI & Reinforcement Learning Engineer. Designing multi-agent self-play systems, deep computer vision, and defensive cyber architectures with tactile clarity and sub-millisecond execution.
Core Engineering Pillars
Autonomous Agents (RL)
Formulating multi-agent self-play systems using PPO with GAE and MADDPG, accelerated across CUDA environments with custom continuous Pymunk physics.
Computer Vision & Security
Pioneering aesthetic full-image QR steganography using Stable Diffusion and ControlNet for defense and distance-authenticated cryptographic payloads at MeitY.
Cloud & Predictive Models
Engineering robust ARIMA time-series forecasting algorithms, deploying auto-scaling endpoints onto scalable IBM Cloud enterprise infrastructure.
Work & Research Trajectory
- Utilized Stable Diffusion and ControlNet (computer vision) to generate aesthetic, full-image blended error-correction QR codes for covert communications and anti-phishing defense.
- Researched and integrated dual-value QR codes with distance-based, cryptographically signed payloads supporting tiered access control and threat modeling.
- Cleaned and structured volatile enterprise datasets, developing robust ARIMA time-series models to generate predictive analytics.
- Deployed containerized model inference endpoints onto the IBM Cloud platform with high scalability.
- Selected and directed an engineering squad of 15+ student developers from 40+ applicants, maintaining a 90% project delivery rate.
- Spearheaded competitive hackathon sprint development and directed technical architecture presentations to expert panels.
Featured Engineering Systems
MARL Tag: 3v3 Pursuit-Evasion Game
PyTorch continuous self-play multi-agent RL system utilizing PPO with GAE. Parallelized over 8 CUDA environments (30M timesteps) with custom 102-ray perception and 95% victory rate.
Multi-Agent Vehicle Fleet Optimizer
Autonomous urban taxi dispatch simulator combining dynamic A* heuristic routing with MADDPG. Reduced passenger wait times by 20% across 100+ concurrent simulated agents.
Student Performance Diagnostic System
Random Forest Classifier with 96% accuracy. Conducted end-to-end data preprocessing (imputation, encoding) and hyperparameter tuning using GridSearchCV (+10% gain).
Automated Video Generation Engine
End-to-end Python pipeline converting structured media data into automated video content using MoviePy and REST APIs, slashing production turnaround time by 90%.
Technical Competencies
Languages & Core Systems
AI & Machine Learning Frameworks
Research Domains
Cloud & Infrastructure
Academic Background
Bachelor of Technology in Information Technology · Manipal University Jaipur (Aug 2022 – May 2026)
Coursework: Data Structures & Algorithms, Operating Systems, RDBMS, Computer Architecture, Data Communications, Web Technologies.
Connect with Aryan Gupta
Available for 2026 full-time engineering and research roles.