Aishwary Kesarwani

AI Researcher

Prayagraj, Uttar Pradesh, India10 mos experience

Key Highlights

2-time ICPC Asia West Regionalist
Expertise in building ML systems for fintech
Passionate about AI-driven financial solutions

Stackforce AI infers this person is a Fintech professional with strong expertise in Machine Learning and Data Analytics.

Contact

Skills

Core Skills

Machine LearningData Analytics

Other Skills

Analytical SkillsDara Structure and AlgorithmsData PipelinesDatabricksDeep LearningDesigning of algorithmsEngineeringFeature EngineeringJavaLLMs (GPT)Model EvaluationObject-Oriented Programming (OOP)PandasPaymentsProblem setting

About

Hi, I’m Aishwary Kesarwani — a final-year engineering student at LNMIIT, passionate about fintech, AI, and engineering data-driven solutions. I’m currently interning as a Risk Engineering Intern at GreenLight Fintech, where I work on solving real-world problems in data modeling, fraud detection, and financial data analysis. Before diving deep into data science, I spent two years in competitive programming, becoming a 2-time ICPC Asia West Regionalist with a top-50 rank — an experience that shaped my analytical thinking and love for solving hard problems. My technical toolkit includes: Python, Pandas, SQL, Databricks Scikit-learn, TensorFlow, Deep Learning Data pipelines, feature engineering, and ML model development LLM, Transformers, Bert I enjoy building intelligent systems that make financial data actionable — from detecting anomalies to predicting risk. I'm especially excited about how AI is transforming fintech.

Experience

Greenlight

Risk Engineering Intern

May 2025 – Present · 10 mos · Bengaluru, Karnataka, India · On-site

Selected for full-time conversion based on strong performance and impact across applied machine learning, LLM systems, and data-driven decisioning.
Technical Contributions & Impact:
Designed, trained, and deployed a transaction risk ML model on Databricks for loss prevention, optimizing precision–recall trade-offs.
Achieved 52% precision and 30% recall, leading to a ~24% reduction in financial loss while introducing only a 0.05% incremental transaction decline, ensuring minimal customer friction.
Worked on feature engineering, model evaluation, threshold tuning, and business-metric alignment for production deployment.
Built LLM-powered analytics pipelines using GPT, Slack APIs, and Databricks to process and summarize high-volume unstructured customer escalation data.
Implemented automated ingestion and parsing of 7-day rolling Slack conversations, generating structured summaries and actionable insights for Product, Strategy, and Engineering teams.
Integrated context augmentation from Google Docs, Confluence, and 5 weeks of historical data using tool-based retrieval (Toons) to improve response relevance and signal quality.
Extended the LLM system to analyze 30 days of customer feedback (suggestions & criticism), enabling product prioritization and leadership decision-making through trend detection and sentiment-driven insights.
Collaborated with cross-functional teams to translate ML outputs into business decisions, focusing on scalability, reliability, and real-world deployment constraints.
Core Technologies:
Python, Machine Learning, Databricks, LLMs (GPT), Retrieval-Augmented Generation (RAG), Slack API, Data Pipelines, Feature Engineering, Model Evaluation, Product Analytics
Thrived in a fast-paced startup environment and greatly enjoyed working from the GreenLight Bengaluru office, collaborating with highly skilled engineers and product leaders while building production-grade ML systems.