Atishay Kasliwal — Software Engineer, production AI and distributed systems.

Research

Applied research, mostly where models meet real constraints.

Two threads: what central bank language actually predicts, and how to deploy medical imaging models without moving patient data.

Areas

  • ActiveNov 2024 — Present

    Financial NLP and central bank communication

    Stony Brook University

    Extracting stance and stance-change from Federal Reserve publications, and testing whether the delta between consecutive releases carries information beyond what is already priced in.

    Read the case study
  • CompletedMay 2025 — Aug 2025

    Deployable ML for medical imaging

    Wake Forest University — Center for Artificial Intelligence Research

    Brain tumor segmentation with a deployment constraint: inference must run client-side so patient imaging never leaves the reviewing machine.

    Read the case study

Publications

  1. Analysis Report on India's Reopening

    Atishay Kasliwal, Saanya Lasod, Mradul Kasliwal

    International Journal for Research Trends and Innovation (IJRTI) · Vol. 5, Issue 6 · pp. 64-69 · June 2020

    Using the dataset from our earlier survey on the same topic, we predicted respondents’ views on the Government of India’s "Unlock" phase 1 using data analysis and machine learning models.

    • COVID-19
    • Machine Learning
    • Data Analysis
    • Lockdown
    • Containment Zone
    PDF ISSN 2456-3315

  2. Survey paper on "Is India ready to reopen?"

    Atishay Kasliwal, Saanya Lasod, Mradul Kasliwal

    International Journal for Research Trends and Innovation (IJRTI) · Vol. 5, Issue 6 · pp. 20-22 · June 2020

    As "Unlock" by the government was just around the corner, we all had one question in our minds: "Is India ready to reopen?" — even when COVID-19 still prevailed. We surveyed around 31 questions on what people observed and experienced during lockdown, then analysed the data and generated trends.

    • COVID-19
    • Lockdown
    • Unlock
    • Survey research
    PDF ISSN 2456-3315

Conference papers

  1. Presented2025

    Market Expectations and FOMC Speech: a Large Language Model Approach to Asset Pricing

    Qianqian Zhang, Wencui Han, Atishay Kasliwal

    International Conference on Business, Management, Economics, and Information Systems · New York, United States

    Applies large language models to Federal Reserve speech to study how market expectations are priced, extending the FOMC research pipeline into an asset-pricing setting.

    Related case study

Graduate coursework

Course presentations

Stony Brook University · MS Data Science

  • Cross-Lingual Document Retrieval using Embedding Models

    AMS 691.02 — Natural Language Processing · Professor Jiawei (Joe) Zhou

    Retrieval across languages using multilingual embedding models, so a query in one language surfaces relevant documents in another.

    With Rutika Avinash Kadam, Sai Ruthvik Madireddy

    • NLP
    • Embeddings
    • Retrieval

  • Cross-Lingual Document Retrieval and Question Answering

    AMS 560 — Big Data Systems, Algorithms and Networks · Professor Zhenhua Liu

    Extends the cross-lingual retrieval work into a question-answering system, breaking the language silos that keep parallel corpora isolated from each other.

    With Meet Zalavadiya, Moraish Kapoor, Shaunak Mahajan, Swati

    • Retrieval
    • Question Answering
    • Distributed Systems

  • Distributed Healthcare Cost Prediction with Spark and PySpark

    AMS 598 — Big Data Analysis · Dr. Song Wu

    Benchmarks Apache Spark against traditional systems for large-scale healthcare cost prediction, and shows that behavioural indicators can predict high drug spending without access to direct cost data.

    With Aishwarya Bhanage, Harsha Vardhan Reddy Vummadi, Sriram Vivek Sanjyot Amritkar, Dhananjay Sharma and 4 others

    • Apache Spark
    • PySpark
    • Distributed Computing
    • Healthcare

  • Forecasting and Segmentation of Guest Booking Patterns in Hospitality

    AMS 597 — Statistical Computing

    Forecasting booking demand and segmenting guest behaviour to understand cancellations and fluctuating demand in the hospitality industry.

    With Venkata Sai Ashrit Kommireddy, Abhishek Sai Maddineni, Meet Zalavadiya, Harish Kandan

    • Forecasting
    • Segmentation
    • Statistical Computing

  • MovieDB: A Comprehensive Movie Database System

    Data Management

    A relational database system for cataloguing films and tracking user activity, covering schema design, normalisation, and query performance.

    With Abhishek Sai Maddineni, S. Varshaa Sai Sripriya

    • PostgreSQL
    • Schema Design
    • Data Modelling