Applied Projects.

Beyond academic research, I build software systems—from interpreters to autonomous driving pipelines—that bridge the gap between theory and execution.

01
Ansh.AI
Deep Dive
HackCU11 (2nd Place AI Track)

Ansh.AI: Conversational Diary

LLM Hackathon

An emotional sounding board and personal logging system. Awarded 2nd place in the AI Track at HackCU11.

02
AWS DeepRacer
Deep Dive
Advanced Robotics

AWS DeepRacer Autonomous Driving

ROS 2 Control

Full ROS 2 autonomous pipeline covering perception (lane detection) and PID control loops.

03
AHRI Project
Research
Research Project

AHRI: Dual Task Modalities

HRI Dual Task

A comparative study on human performance and cognitive load during concurrent dual-task robotic scenarios.

04
Lox Interpreter
Deep Dive
Systems Programming

Lox Language Interpreter

Compilers

Full end-to-end interpreter implementation including scanner, parser, and tree-walk evaluator.

05
V-BIKES
Deep Dive
IoT Safety

V-BIKES Collision Detection

Hardware Sensors

Hardware-based safety system explicitly designed for two-wheeler urban environments.

06
Gaze Estimation
Slides
CVPR Presentation

End-to-end Frame to Gaze Estimation

Vision

Literature review and "End-to-end Frame-to-gaze" CVPR paper presentation for graduate robotics seminar.

07
POMDP Presentation
Slides
Probabilistic Robotics

POMDP Frameworks for Autonomy

Theory

Presentation on Partially Observed Markov Decision Processes and their application in decision-making.

08
DMU Final Paper
Research
Aerospace Engineering

Decision Making under Uncertainty

Autonomy

Research paper on decision-making strategies and algorithms for safety-critical autonomous systems.

09
UNESCO Report
Research
Publication

UNESCO Report on AI & Education

Policy

UNESCO-commissioned report examining the global impact of artificial intelligence on future learning.

10
BEL Industry Presentation
Slides
Applied AI Talk

BEL Industry Workshop: Perception

Industry

Industrial session covering the practical adoption of AI-driven planning pipelines in manufacturing.

11
MRCV Presentation
Slides
Graduate Seminar

Advances in Visual-Language Models

Vision VLM

Explored multimodal fusion, promptable VLM architectures, and evaluation for embodied perception.