Madan Lamichhane, Phd Email & Phone Number

San Diego Metropolitan Area

San Diego, California, United States

Leading machine learning projects for developing fraud prediction models for payment card and retail banking transactions.Designed interpretable machine learning model for the team to identify the reason behind the prediction of the neural network model

Greater San Diego Area

• Key Contributor on designing 2019 Card Not Present (CNP) model. This model detects eCommerce transactions quicker and better in preventing significant dollar loss by catching fraudulent transactions much earlier than the traditional methods. This new technology performs 30% better than traditional fraud detection technology on eCommerce transactions. • Developed blended unsupervised outlier model and supervised regression model for detecting frauds on Online Banking transactions. • Simulated the impact on model performance and score behavior change in production when the European Union is going implement 3DS secure system in eCommerce transactions

Greater San Diego Area

• Developed Neural Network and Adaptive Analytic models to detect fraudulent transactions on Payment Card. This adaptive model has the capability to identify the change in fraudulent behavior and customer regular spending pattern and adapt itself to incorporate this change production environment.• Developed an alternative fraud prediction model and compared it with existing Falcon technology in FICO• Formulated the simulation models to study the impact of the BREXIT in the fraud predictions

South Bend, Indiana Area

Project I: Developed efficient mathematical models for computing electrostatic interactions between multipolar (dipolar or quadrupolar) molecules. Compared energy constants of multipolar crystals evaluated from our newly developed methods with analytical results. This is currently a part of the OpenMD software package which is widely used in the Molecular Dynamics (MD) as well as Monte Carlo (MC) communities.Project II: Studied structural and dynamical properties from newly developed electrostatic methods using molecular dynamic simulations and compared results with the computationally expensive Ewald method.Project III: Developed mathematical models for evaluating correct dielectric properties for dipolar and quadrupolar fluids using three different simulation methods. Currently testing validity of results with the experimental as well as previous simulation results.

United States

Physics Tutor: Helped and guided students for doing problems related to electrostatics, current, and electricity.Lab TA:Helped and guided students for performing experiments in the lab related to mechanics and modern physics.

Memphis, Tn

Project I: Synthesized and characterized rare-earth metal doped BaFe12O19 to enhance its magnetic properties.Project II: Fabricated and characterized hard-soft nano-magnetic composite material using electro-less coating method.

United States

Lab TA: helped and guided students in performing experiment in the lab related to mechanics and modern physics.

Phd, Physics

Developed Computational models to evaluate electrostatic interaction between multipolar molecules in the Molecular Dynamic (MD) and Monte.

Master Of Science (Ms), Physics

Synthesized and characterized magnetic nano-particles.