Ayush

I specialize in data/risk analytics along with data analysis & production ML pipelines. I have successfully built an AI platform serving over 1,000 daily users, effectively bridging the gap between theoretical research and practical implementation. My expertise includes LangChain, big data, tableau dashboarding, multiagent systems, and the latest advancements in neural networks/ML.

Lead Data Analyst / AI Engineer at FlowAI (2024-04 – Present)

AI platform development company focusing on research applications and solutions

• Built production multi-agent AI research platform from scratch, scaling to 1,000+ daily active users including PhD research scholars across physics, mathematics, and computer science domains
• Architected ReAct (Reasoning-Action) framework enabling sequential multi-agent coordination for complex academic research workflows and problem decomposition
• Implemented LangGraph agent orchestration for stateful, iterative research tasks with dynamic tool selection and context management
• Designed and deployed full-stack application on AWS cloud with Python backend, API integrations, MLOps, user authentication, and scalable cloud infrastructure for 1000's of users

Model Development/Analytics Manager at EY (Goldman Sachs Client) (2022-10 – 2024-04)

Consulting firm providing assurance and consulting services

• Implemented LangChain-based RAG pipeline optimizing regulatory text QGA and document summarization, reducing manual review time by 60%
• Applied prompt engineering and in-context learning techniques to fine-tune LLM outputs for financial regulatory compliance use cases
• Gained hands-on experience with transformer architectures (FFN, CNN) for document processing and financial text classification

Quantitative Analyst - Market Risk at JP Morgan Chase (2019-10 – 2022-04)

Leading global financial services firm providing investment banking and financial services

• Built anomaly detection models (Jump and Trend detection) reducing trading limit breaches by 40% through time-series forecasting and statistical pattern recognition, VaR Testing G validation
• Automated stress testing, scenario analysis incl. Fed Stress Capital Buffer (SCB) calculation via Python, directly influencing quarterly hedging strategy and $XM capital allocation decisions
• Applied machine learning techniques to risk modeling, scenario SQL generation, and PGL prediction across derivatives portfolios

Quantitative Researcher at Irage Capital Advisory Pvt. Ltd (2019-01 – 2019-08)

Advisory firm specializing in investment strategies and quantitative research

• Developed high-frequency trading algorithms processing microsecond-level tick data, applying ML for pattern detection and market microstructure modeling

M.Sc. in Physics – IIT Madras (2015-08 – 2017-05)

B.Sc. in Physics – Jamia Millia Islamia (2012-08 – 2015-06)