Skills

• Languages & Libraries: Core (Python, PyTorch, SQL & Cloud (AWS)), Machine Learning (Scikit-learn, Pandas, Numpy, XGBoost, LightGBM), Natural Language Processing (NLTK, SpaCy, Transformers, LangChain), User Interface (Gradio, Streamlit), Spark
• Algorithms & Concepts: Regression (Linear/Logistic), Decision Trees, Random Forest, Gradient Boosting Machine, Clustering, PCA, Neural Networks, Deep Learning, Large Language Models, Prompt Engineering, RAG, Fine-tuning
• Others: LaTeX, MS Office (Word, Excel, PowerPoint), Confluence, Git, Jira, Kanban, Mural
• Expertise: Natural Language Processing / Understanding / Generation
• Domains: FinTech (Financial Services + Technology), Consumer Internet based Products & Customer Analytics
• Relevant Coursework: Technical - Natural Language Processing Specialization (Score >90%), Neural Networks and Deep Learning, Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization (Score>90%), Prompt Engineering, Large Language Models, Cloud - AWS, Machine Learning Engineering, etc.; Soft skills - Learnship Business English Level 10, Creating Effective Presentations, Creative Thinking; Research - Introduction to Research, NPTEL, Score: 87%, Rank: Top 1%, Domain - Mutual Funds, Stocks etc.
• Areas of Interest: Applied NLP, GenAI, ML
• Soft Skills: Critical thinking, Communication, Problem Solving, Self-learning, Resilience, Emotional intelligence
• Leadership: Lead a team of 2 data scientists resulting in a publication in IEEE Big Data conference. Lead an annotation project comprising a team of 98 annotators. Lead multiple groups of 3–4 people resulting in various research paper publications (Venues: LREC-COLING, NTCIR-17, FinNLP@LREC-COLING, FIRE, etc.). Functioned as a primary point of contact for technical paper writing initiatives in Fidelity Investments’ Collaboration & Innovation Track.

Published Industrial Projects

• Automated analysis of customer interaction text to generate customer intent information and hierarchy of customer issues US Patent - 2024
• The Effect of Tweets on the Traded Volume of Crypto-Coins IEEE BigData - 2023
• Dichotomic Pattern Mining Integrated With Constraint Reasoning for Digital Behavior Analysis Frontiers in AI Journal 2022
• Data Driven Content Creation using Statistical and Natural Language Processing Techniques for Financial Domain Financial Narrative Processing Workshop - 2021
• Applying Transfer Learning for Improving Domain-Specific Search Experience Using Query to Question Similarity ACAI-2020

Open Source Academic Research Projects

Inclusive Investing

Making the investment process more inclusive so that even the economically lower strata of the society can avail financial services.

Topics:

• Improving readability of financial texts
  • Skills: Data Curation, Machine Learning, Natural Language Processing, Readability, Transformers
• Improving reach & engagment of financial social media posts
  • Skills: Transformers, Large Language Models (ChatGPT, Claude), Social Media Analytics

Relevant Publications

• “FinRAD: Financial Readability Assessment Dataset - 13,000+ Definitions of Financial Terms for Measuring Readability” in FNP@LREC-2022 (link)
• “Generator-Guided Crowd Reaction Assessment” in TheWebConf (WWW) 2024 (link)

Improved Investing

Improving the journey of investments

Topics:

• Extracting hypernyms of financial terms
  • Skills: Sentence Transformers, Ontology Mining, Natural Language Processing
• Extracting relationship between financial entities
  • Skills: Relation Extraction between Entities, Financial Text Mining, Language Models

Relevant Publications:

• “Term Expansion and FinBERT fine-tuning for Hypernym and Synonym Ranking of Financial Terms” in FinNLP@IJCAI 2021 (link)
• “Learning to Rank Hypernyms of Financial Terms using Semantic Textual Similarity” in SN Computer Science (Springer) 2023 (link)
• “The Mask One At a Time Framework for Detecting the Relationship between Financial Entities” in FIRE 2023 (link)

Impactful (Green) Investing

Considering environmental aspects while investing

Topics:

• Classifying a financail text as Sustainable or Unsustainable
• Detecting Environmental, Social and Governance (ESG) Issues from financial texts
• Identifying ESG impact type
• Identifying ESG impact duration

Relevant Publications:

• “Ranking Environment, Social And Governance Related Concepts And Assessing Sustainability Aspects Of Financial Texts” in IJCAI-ECAI 2022 (link)
• “A low resource framework for Multi-lingual ESG Impact Type Identification” in FinNLP@IJCNLP-AACL 2023 (link)

Informed Investing

Keeping the investors informed and helping them to make data driven decisions

Topics:

• Detecting exaggerated and in-claim numerals from Financial Texts
  • Skills: Pre-trained Language Models (BERT), Machine Learning
• Evaluating the effect of Social Media Posts by Executives on Stock Prices
  • Skills: Social Media Analysis, Deep Learning (LSTM, GRU)
• Evaluating the Rationals of Amateur Investors
  • Skills: Transformers, Ensemble Learning
• Fine-grained Argument Understanding in Financial Texts
  • Skills: Cross Encoders, Pre-trained Language Models

Relevant Publications:

• “LIPI at the NTCIR-16 FinNum-3 Task: Ensembling transformer based models to detect in-claim numerals in Financial Conversation” in NTCIR-16 2022 (link)
• “Evaluating Impact of Social Media Posts by Executives on Stock Prices” in FIRE 2022 (link)
• “LIPI at the FinNLP-2022 ERAI Task: Ensembling Sentence Transformers for Assessing Maximum Possible Profit and Loss from Online Financial Posts” in FinNLP@EMNLP 2022 (link)
• “LIPI at the NTCIR-17 FinArg-1 Task: Using Pre-trained Language Models for Comprehending Financial Arguments” in NTCIR-17 2023 (link)

Indic Investing

Helping Indians to manage their wealth

Topics:

• Financial Argument Analysis in Bengali
  • Skills: Machine Translation, Multi-lingual NLP, Cross Encoders
• Financial Natural Language Processing for Indian Languages
  • Skills: Multi-lingual Natural Language Processing, Transfromers
• Data driven approaches for predicting success of Indian IPOs
  • Skills: Multi-modal Natural Language Processing, Large Language Models (LLMs), Retrieval Augmented Generation (RAG), Fine-tuning LLMs
• Predicting Ratings of Indian IPOs from Red Herring Prospectus
  • Skills: Large Language Models, Small Language Models, Retrieval Augmented Generation (RAG)
• Budget-Assisted Sectoral Impact Ranking
  • Skills: Stock Market Prediction, Text Ranking
• Multi-Modal Indian Earnings Calls Dataset to Predict Stock Prices
  • Skills: Multi-modal Language Models

Relevant Publications:

• “Financial Argument Analysis in Bengali” in FIRE 2023 (link)
• “IndicFinNLP: Financial Natural Language Processing for Indian Languages” in LREC-COLING 2024 (link)
• “Experimenting with Multi-modal Information to Predict Success of Indian IPOs” (link)
• “Predicting Ratings of Indian IPOs from Red Herring Prospectus” (link)
• “BASIR: Budget-Assisted Sectoral Impact Ranking—A Dataset for Sector Identification and Performance Prediction Using Language Models” (link)
• “MiMIC: Multi-Modal Indian Earnings Calls Dataset to Predict Stock Prices” (link)

FinNLP tools

Open sourcing tools for analysing financial texts

Relevant Publications:

• “FinRead: A Transfer Learning Based Tool to Assess Readability of Definitions of Financial Terms” in ICON-2021 (link)
• “Fincat: Financial numeral claim analysis tool” in FinWeb@WWW 2022 (link)
• “Fincat-2: An enhanced Financial Numeral Claim Analysis Tool” in Software Impacts (Elsevier) 2022 (link)
• “FLUEnT: Financial Language Understandability Enhancement Toolkit” in CODS-COMAD 2023 (link)

Doctoral Thesis

Title

Using Computational Linguistics to Demystify Financial Texts

Summary

(link)

Poster

(link)

Video

(link)