Developed an intelligent car recommendation system.
The application uses the Hugging Face transformers library for natural language understanding and provides personalised car recommendations using zero-shot-classification.
Developed an intelligent car recommendation system.
Developed a LSTM model (Recurrent Neural Network) to predict daily stock prices.
The model is built using Google's Tensorflow.
Initially developed and validated the model on Apple stock; extended the framework to predict daily closing price using technical data across a broader set of US equities.
Built and Deployed a restaurant weekly meal planner in Python.
Engineered a takeaway database consisting meal data 150+ restaurants using SQL and dbt. Classified over 350,000 takeaway options into diverse dietary plans like vegan, gluten-free, halal etc. using OpenAI’s GPT-3.5 model, python, FastAPI, GCP python SDKs (google-cloud-bigquery, google-cloud-storage), Docker, GitHub actions, and deployed the API on GCP Linux VM using a microservices architecture. [Private repo, contact for access!]
Developed a model to predict cost of refurbing vehicles using MAPIE (Model Agnostic Prediction Interval Estimator) and Light GBM.
Built a LGBM MAPIE model and a fastapi endpoint which predicts average and maximum refurb costs for retail-able vehicles.
Developed a data quality report and a XGBoost regression model in Python.
Generated a data quality report for a days to sell data for vehicles. This model predicts the number of days it will take to sell a car at a certain selling price. Also performed feature selection and random search hyperparameter optimisation to achieve a RMSE of 0.7 days. [Private repo, contact for access!]
Built and deployed a model monitoring framework in Python.
Developed and embedded a model monitoring framework (consisting of Long Term Model Drift monitoring, Early warning monitoring and alerting) for reporting model drift of 4 supervised machine learning CLTV models (demand, returns, orders, transaction) using sklearn, seaborn packages in python. [Private repo, contact for access!]
Big data analysis with Apache Spark.
Performed big data analysis on tweets using PySpark, Scala, Hive, Impala and mySQL in VMware to monitor unusual activity during football matches.
Developed an advanced spatial analysis package in R.
Built an advanced package in R for spatial analysis. This includes functions such as GetElevation() which calculates the elevation given a lat and long, FindNearest() which looks for the nearest spatial feature, ConvertCoordinates() which converts easting and northing to longitude and latitude and vice versa.
Developed an Accuweather API package in R.
This R package can be used to get historical weather conditions, hourly and daily forecasts for any location in the world. Also includes unit tests.
Exploring Test Driven Development in Python.
This repo is a simple walkthrough of how to apply test driven development in python.
Engineered a T-SQL Database.
This repo consists of a database engineered for a hypothetical online mail company, Insight Mail. Also contains a database schema as well.