Portfolio

Below you can find some of the projects I've done during my work and studies. These projects demonstrate expertise in Reinforcement Learning, Generative AI, Multi-Arm Bandits, and Deep Learning. Opening any of these will take you either to a project report or a GitHub repository.

Contextual Bandits - Offer Recommendations (CORE)
Contextual Bandits - Offer Recommendations (CORE)

Led development of Target's Contextual Offer Recommendation Engine (CORE), a contextual multi-arm bandit system delivering personalized offers to Target Circle members, optimizing for guest engagement including offer adds and redemption.

Reinforcement LearningContextual Multi-Arm BanditsNeural NetworksMatrix FactorizationNNMFDeep LearningPersonalizationRecommendation Systems
Classification with <b>Deep Reinforcement Learning</b> (DDQN algorithm)
Classification with Deep Reinforcement Learning (DDQN algorithm)

Developed a Medical Risk classification model with Costly Features using Deep Reinforcement Learning (DDQN algorithm) achieving above 91% accuracy.

Reinforcement LearningDDQNDeep LearningCostly FeaturesRisk classification
Face Recognition with Anti-Spoof (CNN) on AWS
Face Recognition with Anti-Spoof (CNN) on AWS

Developed a Face recognition system to reduce fraud in virtual customer onboarding with DeepFace embeddings.

Face RecognitionCNNAnti SpoofingdlibDeep LearningAWSLambdaStep FunctionDynamodbDocker
Graph Network Simulation
Graph Network Simulation

Optimised organization hierarchy to reduce cost and increased productivity with different growth strategies using NetworkX and PySim.

NetworkXPySimDI graphSimulationOptimisation
Share 2 Secure :link:, Winner <b>Microsoft Azure<b> Hack :trophy: :tada:

Share to Secure Application for Blockchain Enabled Insurance on Azure Blockchain Workbench.
Features
• Shared view of truth
• Programmable transfers
• Transparency for relevant institutions

BlockchainAzureHackathon
Route optimisation and Lead allocation app on AWS Data Lake
Route optimisation and Lead allocation app on AWS Data Lake

Developed the serverless architecture of Lead allocation and Route optimisation app on AWS Data Lake.

AWSData LakeLambdaStep FunctionDynamodbECSECRAPI GatewayDockerOSRMFargate
PyraDox OCR API :page_with_curl:
PyraDox OCR API :page_with_curl:

PyraDox is a simple tool for document digitization, text extraction and personal information masking with Tesseract-ocr.

OpenCVTesseractDockerImage ProcessingStreamlitAPI
LDA Topic modeling on pending tickets text data
LDA Topic modeling on pending tickets text data

Analysed customer onboarding journey to increase conversion using Natural language processing - gensim LDA topic modeling on pending tickets text data and developed interactive TAT visualization with plotly sankey.

NLPLDATopic modelingPlotlySankeygensim
Meraki - 1st Runner's Up in MishMash Hack :bike: :trophy: :tada:
Meraki - 1st Runner's Up in MishMash Hack :bike: :trophy: :tada:

Fast Route Planning & Route Optimisation with the help of travelling salesman problem (TSP) algorithm.

Route OptimisationOSRMOpenStreetMapDockerHackathon