Vikash Ahir Email & Phone Number

- Trained AI large language models (LLMs)- Tested for deviation in GPT models- Solved coding problems, writing functional and efficient code- Crafted compelling prompts that drove meaningful AI conversations- Wrote robust test cases to confirm code works efficiently and effectively- Evaluated the quality of AI-generated code, including human-readable summaries of rationale- Created and ranked justifications to further reinforcement learning with human feedback- Wrote complex and nuanced code prompts in python that result in a deviation for GPT models

- Built, documented, and ensured quality and robustness of various data pipelines- Partnered with business units to gather, prioritize, and fulfill their information requirements- Adapted existing ETL solutions to ongoing changes in data sources. The data sources include third party APIs events captured on O2E Brands- Produced high-integrity, transparent, and reconcilable data on a regular basis- Monitored and report issues with data quality and creating automated checks on data quality- Integrated data between disparate systems- Automated data-related tasks- Worked collaboratively in an Agile/Scrum team to plan sprints

Remote

- Created a python data pipeline that extracts unstructured blob data from Azure Storage containers and then transforms & preprocesses it for our deep learning model (keras/tf) - Pipeline appends the deep learning model output as a new column to the transformed blob data in a Pandas Series Dataset and uploads it into our Azure File Share Storage while updating Azure Table Storage values for daily iterations- Deployed the above data pipeline and deep learning model to the Cloud as an Azure Function App Trigger to for full autonomy- Implemented Azure Logic Apps on a trigger for automatic emailing and reporting logs- Implemented a Multiclass Naïve Bayes Classifier to predict numerical category codes from a respective categorical text description (one to many)- Preprocessed data for ML algorithms using pandas, numpy and scikit-learn libraries- Created data pipelines to grab unstructured data, preprocess data, and upload data to the cloud- Deployed pipelines and ML models to the cloud

Canada

- Researched and implemented popular open source machine learning models for simple to complex tasks in Jupyter Notebook using Python- Implemented a Multiclass Naïve Bayes Classifier to predict multiple text categories- Implemented a recommendation engine using collaborative filtering with implicit user feedback to predict most popular and potential products for each customers. This would increase new or reoccurring purchases and business acquisition (cold start problem)- Conjoined a gradient boost ML model to my recommendation engine above to predict product prices for each product*customer recommendation pair

- Developed and deployed a secure payment gateway to maintain millions of dollars in business transactions during Covid-19 in a three-week deadline- Managed the whole project start to finish. Researched and implemented the most efficient and upcoming technologies to produce a gateway and procure a multi-million-dollar partnership- Conducted business meetings dealing with BDO Executives and Stripe Sales and Engineering teams in San Francisco- Implemented a custom Microsoft Azure Authorization Server on top of Reacts Next.js framework using JavaScript to secure the gateway for specific employee-only login- Created multiple APIs to connect the front end to Stripes backend in order to customize and allow the portal having monthly payments, one-time billing, and send custom meta-data along with webhooks

Vancouver, Canada Area

Vancouver, Canada Area