Subhasish Chatterjee

• Working for Credit Suisse for their Python and Machine Learning projects for their e-data
• As single contributor created an end-to-end machine learning platform to read requests from mail and convert that to sql queries and run it against the database and return the results in mail to stakeholders. All the checks are placed for error handling.
• Worked on revenue generating model which figures out trade routes for maximum revenue for credit suisse.
• Carried out day-to-day duties accurately and efficiently.
• Worked on azure databricks prototype for time series anomaly detection.
• GCP ML and Data Engineering prototypes

Role:

• Leading a team of 6 members including Python Developers, Big data developers and Machine Learning Engineers Managing Machine Learning and Deep Learning Models covering feature engineering, feature selection, model validation
• Using machine learning and statistical modelling techniques to develop and evaluate algorithms to improve performance, quality, data management and accuracy
• Working in Innovation Lab to make the state-of-the-art in research centric products and features using Data Science/AI otherwise known as accelerator programs
• Designing lean proofs of concepts (POC) to answer targeted business questions
• Building new models arounds different business problems like risk analytics, NLP pretrained models, financial forecasting, portfolio optimization, seq2seq models.

Highlights:

• Received Surpass Award for outstanding performance: For the period Jan to March 2019, conferred by CEO of Synechron, Faisal Husain
• Won Hackathon, been invited to Insync event, New York , April 18, 2019 at Mandarin Oriental, New York
• Project: Accelerator Programs, Financial Modelling, NLP Requirements, Quant
• Project: Neural Search Query | Client: CITIBANK
• Description: Building a ground up tool for management and stakeholders and query different data sources i.e. Elastic Search or APIs using natural language and receive the data in matter of seconds completely bypassing the database complexities of fetching a data. The project is multifaceted with features including recommender systems, text summarization, text generation, training data automation among many things.
• Project: Model Management Framework | Client: Wells Fargo
• Description: A project that started with the needs of Wells Fargo Quant teams move to switch from SAS to Python. A framework should have the provision of feature engineering, feature selection, Model tracking, and automated training of financial models among many things. These mostly done from the prospects of Risk Analytics hence datasets like Fannie Mae, Lending Club were heavily used.

• Evaluated diverse organizational systems to identify workflow, communication and resource utilization issues.
• Compiled research data and gave professional presentations highlighting finds and recommended optimizations.
• An AI Powered Chatbot capable of client interaction.
• Prototype Language generation tasks for internal use

• Developed diagrams to describe and lay out logical operational steps.
• Presented business leaders with ROI estimations to drive project planning on recommender systems for better online user interaction
• Creating regression analysis to investigate system issues and implemented resolutions to reduce downtime.
• Implement Information Systems checks and balances.