Headline

Serving Notice Period | AI/ML Engineer BNY Mellon | IIT Madras M.Tech AI | GenAI · MultiAgent · RAG · LLM Fine-tuning | Production AI in fintech | Contributor at Analytics Vidhya, Medium | Expert codeforces | 5⭐️codechef

About

ML Engineer with 3+ years building production AI systems in financial services. I build things real users depend on daily production systems at scale. What I've shipped: → Multi-agent MCP platform orchestrating 4-5 AI agents across multiple data sources (CrewAI/A2A) → NL-to-Elasticsearch engine serving 800-1000 wealth managers daily (GPT-4o + BERT embeddings) → LLM fine-tuning pipeline - LoRA on Phi-3 mini, ~95% accuracy on domain tasks → RAG systems for wealth management document retrieval → STT pipeline for financial call transcription and analysis I also write about how production AI systems work under the hood from reverse-engineering Gemini's memory architecture to building custom evaluation metrics for LLM fine-tuning with LLaMA-Factory. Links in Featured. Currently pursuing M.Tech in Artificial Intelligence from IIT Madras. Competitive programming: Codeforces Expert | CodeChef 5-Star Stack: Python, PyTorch, LangGraph, CrewAI, MCP, RAG, LoRA, Elasticsearch, FastAPI, Docker Open to ML Engineer and Applied Scientist roles at companies building AI at scale.

Location

Pune Division

AI stack

antigravity

claudecode

Work experience

Artificial Intelligence Engineer at BNY, Jan 2024 - Present · 2 yrs 8 mos, Fine-tuning LLMs (Phi-3 Mini) for Finance: Fine-tuned Phi-3 Mini with LoRA, enhancing financial query resolution in internal AI assistants. Multi agentic system based on MCP. NLP-to-ElasticSearch Query Converter: Developed a GenAI-powered system for converting natural language queries into optimized ElasticSearch DSL queries, improving search accuracy and accessibility. AI-driven Regression Test Prioritization: Built an AI-based test selection system using LSTM models and vector embeddings, reducing test execution time. Embedding-Based Similarity Search: Designed an optimized search engine for financial documents using FAISS. Model Optimization: Worked on LoRA tuning, quantization, and retrieval-augmented generation (RAG) for improving GenAI efficiency.

Software Engineer at BNY, Jul 2023 - Jan 2024 · 7 mos, Pune District, Maharashtra, India, Enterprise-Scale Software Control Management Platform: Developed a deployment automation system using App Engine, enabling users to deploy multiple applications via a UI. Optimized deployment time from 2 minutes to 54 seconds through parallel batch processing. Tech Stack: Python, Flask, Google App Engine, Multithreading, Server-Sent Events, Closure Built a WebSocket-based Stock Price Ticker: Developed a NestJS playground with WebSockets to display real-time stock prices based on ticker values. Tech Stack: NestJS, WebSockets Kafka Management Tool Development: Built a Kafka admin tool for creating, deleting topics, and managing user permissions using Golang (goroutines). Tech Stack: Golang, Kafka, Goroutines Python Utility Development: Developed various Python utilities for handling CSV files, S3 storage (using Boto3), email automation, and other backend operations. Tech Stack: Python, Boto3, S3, CSV Processing

SDE Intern at BNY, Jan 2023 - Jun 2023 · 6 mos, Pune District, Maharashtra, India, Contributed to Wove, a one-stop trading platform that suggests investment models to advisors based on client requirements using a rebalancer system. Mail Processing & Notification System: Built an email automation system for handling trading notifications and updates. Kafka Messaging System Development: Designed and implemented Java stubs for testing, allowing seamless message publishing and subscription to Kafka topics for real-time event processing. Tech Stack: Java, Spring Boot, MongoDB, Lombok, Swagger, PySpark, JMS, Kafka

problem curator at iMocha, Jul 2021 - Jan 2023 · 1 yr 7 mos

Education

Master of Technology - MTech, Artificial Intelligence, Indian Institute of Technology, Madras

Bachelor of Engineering - BE, Computer Engineering, Pune Institute of Computer Technology

Public links

github: https://github.com/rushi-02

medium: https://medium.com/@rushikeshchavan_99600

codeforces: https://codeforces.com/profile/rushi_30