Fidelity Investments

Corporate Actions Announcement Summarisation
* Summarisation of Key insights present in a Corporate Actions Announcement document that reduces manual touch points
* Leveraging Multi-query RAG based question answer generation that is used to extract factual information. Curation of RAG into sections to generate topic based summaries 

Domain-specific Question Answering System using LLM
* Leveraging open source Large Language Models to build a domain specific question answering system fine-tuned on Earnings Call Transcripts of select firms across several years.
* Implementation of Retriever-Generator model that searches from a corpus of documents through LangChain, Sentence transformers and FlanT5 fine-tuned using QLORA

Customer Service Optimization in Transfer of Assets
-> Aiming to optimize initial action days in the space of Transfer of Assets (TOA)  using machine learning techniques. 
-> By predicting optimal action days, targetting a 30% improvement in follow-ups, reducing customer churn, and enhancing efficiency.

Next Best Action for T+1 Trade Reconciliation Process
* Analysis of trade break life cycle in the trade reconciliation process.
* Incorporation of the next best action model that helps in the break resolution which results in a significant reduction in break cycle time facilitating the industry to move towards T+1 settlement cycle.

Data Science Evangelist
* Roadshow of Artificial Intelligence and its capabilities on financial operations to the Senior Management Team. 
* Curated a backlog of potential business AI/ML use cases in the spaces of Risk Management, Document processing, Anomaly detection, Brokerage operations, Reconciliation, and Settlements

Lead - Data Scientist

Customer Margin Unsecured Deficit Account Classification
- Enhancements to the deficit classification problem through a multi-class classification machine learning model. 
- This reduced the weekly reserve calculation to a daily process resulting in 100 hours of effort avoidance per month and a 50% reduction to the total margin reserve amount. 
- Functioned as the technical lead of the initiative by facilitating the end-to-end delivery of the use case.

Customer Portfolio Anomaly Detection
- Improvisation of the anomaly identification process on customer accounts and reduction of the manual review process by 90%. 
- Leveraged stock index data and several statistical methods processing millions of records daily to identify the current market trend to effectively identify and flag potential anomalies through machine learning.

Transformer-based Search Retrieval Application
- Implementation of a search application with capabilities to retrieve the relevant results semantically from a document corpus via custom-trained BERT and an entity identification model to identify stock tickers, CUSIPs, and key phrases through NER model.

AI/ML SE - Data Scientist

Intern

As an experienced Lead - Data Scientist with a passion for applying Machine Learning and Natural Language Processing to solve real-world business problems, I have spent over 4 years developing insightful data-driven solutions in multiple business functions within the financial domain including Financial operations, Risk mitigation, Reconciliation, and, Anomaly detection.

Over the past 2 years, I have had the privilege of leading projects end to end, taking charge of various data science initiatives and successfully delivering comprehensive machine learning solutions.

With my expertise in Applied machine learning, Data Visualization, Statistical Modeling, Predictive Analytics, and Technical Leadership, continuing to enjoy the journey by solving data science problems and delivering value at scale.

Additionally, as a Data Science Evangelist, I'm passionate about educating the capabilities of Machine Learning in Business operations, enabling organizations to harness the power of data for success.

Taplio

Abilash B's Linkedin Analytics

Abilash B

Want to drive more opportunities from LinkedIn?

Taplio

.css-1jgreyo{background-image:linear-gradient(284deg, rgba(39,43,72,1) 0%, rgba(33,47,183,1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Abilash B's Linkedin Analytics

Abilash B

Want to drive more opportunities from LinkedIn?

Abilash B's Linkedin Analytics