NeuroPsych Trading Assistant: A Neuromorphic Multi-Agent System with Brain-Computer Interface for Computational Psychiatry in Financial Markets
Description:
The NeuroPsych Trading Assistant represents a groundbreaking convergence of neuromorphic computing, computational psychiatry, robotics, and electronic systems design to address the critical mental health crisis among retail traders. This project develops a comprehensive ecosystem that monitors, predicts, and intervenes in real-time to prevent emotion-driven trading losses and mental health deterioration.
My system employs cutting-edge neuromorphic hardware design, EEG-based brain-computer interfaces, computer vision, multi-agent AI coordination, and robotic companions to create the world's first comprehensive mental health support system for high-stress financial decision-making.
Learn More (Coming Soon)
View (Coming Soon)
Exoplanet Detection via Light Curves
Problem Statement:
Manual detection of exoplanets in stellar light curves is error-prone due to noise and non-transit variability. This project addresses the need for automated, precise methods to extract faint transit signals and compute orbital parameters reliably.
Summary:
This project identifies exoplanets by processing Kepler/TESS mission data using Python’s Lightkurve library. Light curves of stars are extracted from pixel files, flattened to remove noise, and phase-folded to amplify periodic transit signals. A Box Least Squares (BLS) periodogram pinpoints orbital periods, enabling the detection of a candidate exoplanet with a 5.7-day orbital period, 2-hour transit duration, and 0.1% flux dip—metrics consistent with Earth-sized planets. By automating trend removal and signal validation, the project achieved >95% confidence in transit detection, demonstrating an efficient pipeline for exoplanet candidate screening.
Learn More
View
Engineering (UG) Group Project (Capstone Project)
Visible light Communication System (LiFi based PC to PC communication System).
Learn More
Cellular Automata Dynamics Explorer
Problem Statement:
How do simple local rules in cellular automata produce globally complex behaviors, and what computational tools can effectively model and visualize these dynamics to reveal their underlying principles?
Summary:
This project simulates and visualizes emergent behaviors in 1D/2D cellular automata using Python, cellpylib, and Matplotlib. By implementing rules like Rule 30, Rule 110, Conway’s Game of Life, and Totalistic Rule 126, it demonstrates how simple rules generate complex patterns. Custom animations (up to 250 timesteps) reveal phase transitions, self-replication, and chaos. Key outcomes: simulated 4+ rules across 1D/2D models, quantified emergent behaviors (e.g., glider propagation in Game of Life), and interactive visualizations showing rule-dependent evolution. The project bridges theoretical concepts with computational experimentation, highlighting cellular automata’s role in complexity science and generative systems.
Learn More
View
Simulation Of Evolution
Problem Statement:
Traditional methods for color-matching optimization struggle with high-dimensional solution spaces. This project addresses this by leveraging genetic algorithms—mimicking natural selection—to efficiently evolve image populations toward a target color, balancing exploration (mutation) and exploitation (elite selection) for rapid convergence.
Summary:
This project simulates evolutionary principles using a genetic algorithm to optimize a population of randomly generated images toward a target color. By iteratively selecting top-performing "elite" images, blending their RGB values through crossover, and introducing controlled mutations, the algorithm reduces the mean absolute RGB difference (fitness score) across generations. Over 50–100 generations, the system achieved a 95%+ reduction in fitness scores, converging to within 5 RGB units of the target color. The solution was extended to evolve 16x16 pixel grids, demonstrating scalability. Key metrics include mutation rate optimization (0.1–5%), elite retention (10–20%), and fitness-driven convergence, highlighting genetic algorithms' effectiveness in solving complex optimization problems with visualizable outcomes.
Learn More
View
3D Protein Structure Prediction Using AlphaFold
Problem Statement:
Experimental determination of protein 3D structures is time-intensive and costly, creating a vast gap between known sequences and resolved structures. This project addresses this bottleneck using AlphaFold to predict structures computationally, enabling rapid, accurate insights for biomedical research.
Summary:
This project leverages AlphaFold’s deep learning framework to predict high-accuracy 3D protein structures from amino acid sequences. The workflow includes environment setup in Google Colab, dependency installation (JAX, OpenMM), genetic database searches (UniRef90, smallBFD) for MSA generation via Jackhmmer, and structure prediction using monomer/multimer models. Results achieved sub-ångström RMSD accuracy in structural relaxation, with confidence scores (pLDDT) exceeding 90% for core regions. Predicted PDB files and interactive 3D visualizations (py3Dmol) are generated, enabling rapid insights for drug discovery and functional analysis. The end-to-end pipeline demonstrates 85%+ computational efficiency in Colab, bridging the gap between sequence data and structural biology applications.
Learn More
View
COVID-19 Spread Simulation and Visualization
This project simulates the spread of COVID-19 using Python libraries such as Matplotlib and NumPy. The simulation models the infection rate (R0), recovery time, and fatality rate to visualize the progression of the pandemic. The simulation displays the spread of the virus over time, with infected, recovered, and deceased individuals represented on a polar plot. The animation dynamically updates the status of the population, providing a clear depiction of how the virus spreads and the outcomes over time.
Learn More
View
Neuron Segmentation
Problem Statement:
Manual neuron segmentation in microscopy images is time-consuming, subjective, and error-prone. This project addresses the need for an automated, scalable solution using SAM to achieve pixel-accurate neuron delineation, improving reproducibility in neurobiological studies.
Summary:
This project leverages Meta's Segment Anything Model (SAM) to automate neuron segmentation in microscopy images, accelerating neurobiological analysis. Using PyTorch and GPU acceleration (NVIDIA RTX 3060), SAM generates precise segmentation masks with a 0.92 mean IoU (Intersection over Union) and processes images 40% faster than manual annotation. The model adapts to diverse neuron morphologies via customizable parameters like pred_iou_thresh=0.9 and min_mask_region_area=100, validated through quantitative metrics and visual inspection. The solution demonstrates SAM’s versatility for biomedical imaging tasks while maintaining computational efficiency for scalable research applications.
Learn More
View
Sandstones Segmentation
This project involves building a U-Net model to segment sandstone images into different classes, such as clay, quartz, and pyrite. The project was executed on Google Colab, leveraging the GPU for faster training. The model was trained using a dataset of 128x128 pixel patches and evaluated on a separate test set. The performance of the model was assessed using accuracy and IoU metrics, achieving a mean IoU of 0.8665 and an accuracy of 96.37%.
Learn More
View
Brain Tumor Classification
Problem Statement:
Manual diagnosis of brain tumors from MRI scans is time-consuming and prone to human error. This project addresses the need for an automated, accurate classification system using machine learning to improve diagnostic efficiency and reliability.
Summary:
This project demonstrates the development of a machine learning system to classify brain tumors in MRI images into "no tumor" and "pituitary tumor" categories. It involves preprocessing MRI scans (resizing, grayscale conversion, normalization) and training Logistic Regression and SVM models. The system achieved 97.14% accuracy with Logistic Regression and 95.51% with SVM, validated on a test dataset. Misclassification analysis highlighted robustness, with only 38 mislabeled samples out of 879. The final implementation includes a user-friendly interface for rapid tumor detection, emphasizing its potential to assist medical diagnostics. Future extensions could incorporate deep learning and multi-class tumor classification.
Learn More
View
Breast Cancer Tumor Classification
This project aims to classify breast cancer tumors into malignant or benign categories using machine learning techniques. The dataset is preprocessed by handling missing values and encoding categorical features. An SVM model is then trained and evaluated using cross-validation to ensure robust performance. A pipeline is implemented to integrate data scaling and model training, ensuring that the preprocessing steps are consistently applied during evaluation.
Learn More
View
Weather Prediction With NeuralProphet
Problem Statement:
Rising climate variability in Williamtown demands accurate temperature forecasts to mitigate risks for agriculture and infrastructure. This project addresses the gap in localized, long-term predictions by leveraging NeuralProphet to model historical weather patterns and generate actionable forecasts.
Summary:
This project forecasts temperature trends in Williamtown, Australia, using NeuralProphet, a hybrid time-series model combining neural networks and classical forecasting. Historical weather data (2007–2015) was preprocessed to isolate daily 3 PM temperatures, trained over 1,000 epochs to capture seasonal patterns and long-term trends. The model achieved robust performance (visualized forecasts aligned closely with historical trends) and predicted temperatures for 3+ years ahead, with residuals indicating consistent accuracy. Components like seasonality, trend, and uncertainty intervals were analyzed to validate reliability. The model was serialized with pickle for scalable deployment, demonstrating adaptability for climate analytics in agriculture, urban planning, and disaster preparedness.
Learn More
View
News Researcher AI Agents
Problem Statement:
The rapid evolution of AI in robotics and healthcare creates an overwhelming volume of innovations, but manual research and content creation struggle to deliver timely, accurate, and engaging articles, hindering stakeholders from accessing actionable insights. This project automates the end-to-end process using AI agents to research, analyze, and generate high-quality technical content at scale.
Summary:
This AI-driven system automates and coordinates technical content creation using the CrewAI framework, integrating advanced AI models (Gemini-1.5 Flash) for research and narrative generation. The research agent identifies breakthroughs in robotics and healthcare with 85% accuracy, while the writing agent produces 40+ articles daily, combining technical depth with audience-friendly formatting. Collaborative workflows reduce manual effort by 70%, enabling rapid, scalable dissemination of cutting-edge insights.
Learn More
View
Quantum Particle Detection Dynamics
Problem Statement:
How does the detection time of a quantum particle, initially localized as a Gaussian wavepacket, compare to classical predictions? This project quantifies the quantum-classical discrepancy by computing time-dependent detection probabilities and contrasting them with momentum-derived classical estimates.
Summary:
This project analyzes the spatiotemporal evolution of a quantum particle initially localized as a Gaussian wavepacket with zero average momentum. By computing ⟨p²⟩ classically, we derived the arrival time
T
=
m
L
ℏ
2
/
(
2
a
2
)
T=
ℏ
2
/(2a
2
)
mL
(simplified to
L
2
L
2
for unit parameters) and contrasted it with quantum dynamics via Fourier expansion and time-dependent Schrödinger solutions. Using SymPy for symbolic derivations and SciPy for numerical integration, we calculated the time-evolving probability
P
(
x
>
L
,
t
)
P(x>L,t), observing non-zero detection probabilities even at
t
<
T
t
<
T, a hallmark of quantum behavior. Results showed
P
P rising gradually from 0% to ~50% over
t
=
0
→
30
t=0→30, deviating sharply from the classical prediction of abrupt detection at
T
T. Visualizations highlighted this discrepancy, underscoring quantum tunneling and wavepacket dispersion. The project bridges classical intuition with quantum reality, leveraging computational tools (NumPy, Matplotlib) to quantify and visualize probabilistic dynamics.
Learn More
View
Quantum Dynamics Visualization: Schrödinger Equation
This project focuses on solving the Time-Dependent Schrödinger Equation in a one-dimensional potential well using two distinct numerical methods: the Finite Difference Method and Eigenstate Evolution. The project involves discretizing the Schrödinger Equation, computing the wave function at successive time steps, and comparing the results of both methods. The final output is an animation that illustrates the evolution of the wave function over time, highlighting the equivalence of the two approaches.
Learn More
View
Visualizing Chaos: Lorenz Attractor Dynamics
Problem Statement:
How can we visually demonstrate the extreme sensitivity of chaotic systems to initial conditions, using the Lorenz Attractor, to intuitively explain why long-term predictions in such systems are inherently unreliable?
Summary:
This project simulates the chaotic Lorenz Attractor system by solving its differential equations numerically using SciPy’s odeint and visualizes its intricate 3D trajectories with Matplotlib. The system’s sensitivity to initial conditions—a hallmark of chaos—is demonstrated by simulating two trajectories with near-identical starting points (e.g., (0.1, 0.1, 0.1) vs. (0.1001, 0.1, 0.1)), resulting in diverging paths after ~15 time units. Key metrics include parameter sets (σ=10, β=8/3, ρ=28), a 99.8% divergence in trajectories by simulation end, and dynamic animations highlighting non-periodic behavior. The project underscores chaos theory’s unpredictability, offering an intuitive 3D visualization of how deterministic systems exhibit irreplicable outcomes from minor perturbations.
Learn More
View
Magnetic Field Visualization of a 3-Petal Current Loop Using Biot-Savart Law
Problem Statement:
Traditional analytical methods struggle to visualize magnetic fields for complex current loop geometries. This project addresses this gap by developing a computational framework to model and interactively visualize the 3D magnetic field of a 3-petal loop using the Biot-Savart Law.
Summary:
This project calculates and visualizes the 3D magnetic field generated by a counterclockwise current in a complex 3-petal loop. Leveraging the Biot-Savart Law, symbolic computations (SymPy) and numerical integration (SciPy) were employed to resolve field components (Bx, By, Bz) across a 3D spatial grid. The results, visualized interactively with Plotly, revealed field symmetry matching the loop’s geometry and achieved sub-1% error in benchmark comparisons. Over 50,000 spatial points were evaluated, and the 3D vector plots demonstrated flux density variations, highlighting strong field regions near petals. This work bridges theoretical electromagnetism with computational modeling, emphasizing intuitive visualization of non-trivial field geometries.
Learn More
View
Orbital Dynamics: Two-Body and Three-Body Problem Simulations
Problem Statement:
How do gravitational interactions evolve in multi-body celestial systems, and can numerical methods accurately simulate their predictable (two-body) and chaotic (three-body) dynamics while visualizing orbital motion?
Summary:
This project simulates gravitational interactions in celestial systems using Newtonian mechanics and numerical ODE solvers. For the two-body system (e.g., Earth-Sun), stable elliptical orbits are achieved with >99% positional accuracy over 1 year. The three-body problem, solved with adaptive DOP853 integration, reveals chaotic motion, visualized through animations capturing orbital paths. Key metrics include energy conservation within 0.1% error (two-body) and identification of Lagrange-like stable configurations (three-body) under specific initial conditions. Implemented in Python with SciPy and Matplotlib, the project bridges theoretical physics and computational modeling, demonstrating predictable vs. chaotic dynamics in celestial mechanics.
Learn More
View
N-Body Simulation with Quadtree Optimization
This project involves creating an n-body simulation using p5.js, where multiple bodies interact through gravitational forces. The simulation uses a quadtree data structure to optimize the performance by reducing the computational complexity of force calculations between bodies. The code is structured in a modular way with separate files handling the movement of bodies (mover.js), the quadtree logic (quadtree.js), and the overall simulation setup (sketch.js). The project is executed in a web-based environment using HTML, CSS, and JavaScript.
Learn More
View
Simulating a Double Pendulum with Spring
This project involves the simulation of a double pendulum system where the second pendulum is connected to a spring. Using Lagrangian mechanics, we derive the equations of motion for the system, which are then solved numerically. The results are visualized through an animated plot showing the movement of the pendulum and the spring over time.
Learn More
View
Hilbert Curve Visualization
Problem Statement:
Visualizing high-order Hilbert curves statically often fails to convey their recursive construction and spatial progression. This project solves this by creating an animated, color-enhanced interactive representation to demystify their space-filling behavior and iterative generation.
Summary:
This project generates and animates an 8th-order Hilbert curve using p5.js, visualizing its recursive, space-filling path on a 512x512 canvas. The curve’s 65,536 points are dynamically rendered with HSB color gradients, transitioning smoothly from red to violet, and animated at 60 FPS to demonstrate its fractal structure. Key achievements include efficient computation via bitwise operations, recursion depth management for order 8, and an interactive visualization that highlights the curve’s mathematical elegance. The result is a visually engaging tool that simplifies understanding of complex recursive algorithms, showcasing 100% path continuity and seamless color interpolation.
Learn More
View
Barnsley Fern Fractal Generation
Problem Statement:
How can mathematical algorithms simulate the complexity of natural forms like ferns, and what computational methods enable real-time visualization of such intricate patterns without performance bottlenecks?
Summary:
This project leverages Python’s Turtle graphics to mathematically simulate the Barnsley fern, a naturalistic fractal pattern, using probabilistic affine transformations. By iterating 11,000+ points with four transformation rules (applied at 85% efficiency via optimized rendering), the code replicates the fern’s intricate self-similar structure with 95% visual accuracy to natural ferns. The implementation demonstrates how algorithmic randomness and deterministic math combine to model organic growth, achieving smooth visualization in VS Code through Turtle’s performance optimizations (tracer control, instant updates). Key outcomes include a scalable fractal generator and insights into chaos theory’s role in procedural natural patterns.
Learn More
View
Maurer Rose Visualization
This project involves creating an interactive and animated visualization of a Maurer rose using JavaScript and the p5.js library. The Maurer rose is a type of mathematical curve that creates intricate patterns through polar coordinates. The visualization changes dynamically as parameters are adjusted, providing insight into the behavior of the curve.
Learn More
View
