SAYANTAN ROY

AI Cloud Engineer | Software Developer (AI/ML)
Bengaluru, IN.

About

Highly accomplished AI Cloud Engineer and Software Developer with over 2 years of experience in designing, developing, and deploying cutting-edge AI/ML solutions and scalable cloud infrastructure. Proven expertise in machine learning, deep learning, NLP, and MLOps, with a strong track record of building high-performance systems, optimizing server resources, and integrating complex payment gateways. Adept at leveraging advanced technologies like LLMs, Kubernetes, and PyTorch to drive significant business impact and enhance operational efficiency.

Work

Krutrim
|

AI Cloud Engineer

Summary

Leading cloud engineering initiatives to build scalable AI infrastructure and optimize system performance.

Highlights

Engineered and integrated FusionAuth with Krutrim Management Dashboard, establishing robust Role-Based Access Control (RBAC) across diverse organizational structures to enhance security and access management.

Orchestrated the deployment of a comprehensive Observability Stack, leveraging Grafana-Postgres HA in Kubernetes, to provide real-time monitoring and performance insights for K8s, OpenStack, and Ceph clusters, utilizing VictoriaMetrics for data collection.

Optimized open-source Large Language Models (LLMs) via LoRA on a dataset of 5,000 Jira tickets, resulting in the development of an AI-powered chatbot that automates and resolves L1 team support queries, improving efficiency.

Teliolabs
|

Software Developer (AI/ML)

Summary

Developed and deployed advanced AI/ML solutions, including chatbots, recommendation engines, and forecasting models, enhancing product capabilities and operational efficiency.

Highlights

Architected and developed a Generative AI chatbot incorporating conversational memory and Retrieval Augmented Generation (RAG) strategy, utilizing React-RTK for the frontend and Python Fast API for the backend.

Designed and implemented a high-accuracy recommendation engine using Transformers and PyTorch, achieving 87% prediction accuracy for auto-completing financial queries by training on a proprietary 50K+ Fintech dataset.

Engineered Time-Series Forecasting models that achieved 92% prediction accuracy for document upload and download patterns, significantly optimizing server resource allocation and efficiency.

Seamlessly integrated Stripe Payment Gateway and Stripe Connect APIs, streamlining and securing fund transfers across diverse user accounts.

Publications

Reddit Comment Toxicity Score Prediction through BERT via Transformer Based Architecture

Published by

2022 IEEE 13th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON)

Summary

Published research on predicting Reddit comment toxicity scores using a BERT-based Transformer architecture, presented at the 13th IEEE IEMCON.

Education

IIIT, Naya Raipur

Bachelor of Engineering

Data Science and Artificial Intelligence

Grade: 8.56 CGPA

Courses

Machine Learning Algorithms

Deep Learning Architectures

Natural Language Processing

Cloud Computing Fundamentals

Data Structures & Algorithms

Database Management Systems

Artificial Intelligence Principles

Big Data Technologies

Awards

1st Position at IAM (Industry Academia Meet) in Data Science and ML Track

Awarded By

IIIT, Naya Raipur

Achieved first place in the Data Science and ML Track at the Industry Academia Meet.

Prize for Filecoin Track at HackForTomorrow

Awarded By

HackForTomorrow

Recognized with a prize in the Filecoin track at the HackForTomorrow event.

Top 50 Teams in Shell.ai Hackathon for Sustainable and Affordable Energy 2021

Awarded By

Shell.ai Hackathon

Ranked among the top 50 teams in the Shell.ai Hackathon for Sustainable and Affordable Energy.

Cleared JEE Advance with a Rank of 20k

Awarded By

JEE Advance

Successfully cleared the Joint Entrance Examination (JEE) Advance with a rank of 20,000.

Projects

Two Tower Recommendation System

Summary

Implemented a deep learning recommendation system using a Two-Tower neural architecture in PyTorch, incorporating negative sampling techniques. Built custom dataset handlers and efficient training pipelines that process user demographics and item metadata to deliver personalized Top-N recommendations through vector similarity matching.

Skills

TypeScript

Programming Language, Frontend Development.

Python

Programming Language, Data Science, Machine Learning.

Java

Programming Language, Backend Development.

JavaScript

Programming Language, Frontend Development.

Solidity

Blockchain, Smart Contracts.

Next Js

Frontend Framework, React.

ExpressJs

Backend Framework, Node.js.

Tensorflow

Machine Learning Framework, Deep Learning.

PyTorch

Machine Learning Framework, Deep Learning.

Langchain

LLM Development, AI Framework.

Scikit-Learn

Machine Learning Library, Data Analysis.

Pandas

Data Manipulation, Data Analysis.

Kafka

Distributed Messaging, Data Streaming.

Spark

Big Data Processing, Distributed Computing.

Postgres SQL

Relational Database, SQL.

NoSQL

Database, MongoDB, Cassandra.

VectorDb

Vector Database, Embeddings.

Machine Learning

AI, Algorithms, Predictive Modeling.

Deep Learning

Neural Networks, CNN, RNN, Transformers.

NLP (Natural Language Processing)

Text Analysis, Language Models.

LLMs (Large Language Models)

Generative AI, Fine-tuning.

RAG (Retrieval Augmented Generation)

Generative AI, Information Retrieval.

AWS

Cloud Computing, Cloud Services.

Lambda

Serverless Computing, AWS Lambda.

Azure

Cloud Computing, Microsoft Azure.

Docker

Containerization, DevOps.

Ngnix

Web Server, Reverse Proxy.

Kubernetes

Container Orchestration, DevOps.