Summary
Overview
Work History
Education
Skills
Certification
Projects
Languages
Timeline
Generic

Koyana Dutta

Bally

Summary

AI Engineer with 2 years of experience in NLP, LLM integration, and predictive analytics. Skilled in designing scalable ML solutions, building custom chatbots using Rasa and LangChain, and deploying applications with FastAPI and Docker. Proven ability to improve model accuracy and automate workflows across domains including e-commerce, HR tech, and image generation.

Overview

3
3
years of professional experience
1
1
Certification

Work History

AI Engineer

CLAPPIT Pvt. Ltd.
03.2024 - Current
  • Integrated RAG pipelines with Gemma-9B and Mistral-7B for advanced NLP-based question answering
  • Built a secure Rasa chatbot with FastAPI backend, improving data security and reducing latency by 35%
  • Developed an intelligent SQL agent using DSPy and Microsoft Phi-3, boosting query generation efficiency by 60%
  • Created a Gemini AI-powered chat interface using Langchain and Streamlit, improving response accuracy by 70%
  • Implemented image generation workflows using vision-language models for real-time user inputs

AI Consultant

CLAPPIT Pvt. Ltd.
12.2023 - 02.2024
  • Integrated Kubernetes clusters with K8sGPT and Kafka for real-time monitoring and command response
  • Developed custom chatbots leveraging LLMs and explored RAG-based search with Llama 2 for domain-specific applications

Data Scientist - Trainee

One Billion Ideas Pvt. Ltd.
04.2023 - 02.2024
  • E-commerce Analytics: Processed 10K+ daily transactions, increasing average order value by 12% and customer satisfaction by 20%
  • Computer Vision: Developed Raspberry Pi-based face detection using CNNs, reducing false negatives by 50%
  • Conversational AI: Built chatbot using Langchain and Claude, cutting support costs by 50%
  • Voice Recognition: Developed a voice-to-text system using Whisper, achieving 95% transcription accuracy

Data Science Intern

One Billion Ideas Pvt. Ltd.
01.2023 - 03.2023
  • Conducted anomaly detection for stock trend forecasting and created Flask APIs for streamlined data pipelines

Data Science & ML Intern

Gilbert Research Center
10.2022 - 10.2022
  • Contributed to CNN-based signal/image processing by building PCA-driven visualizations for biomedical datasets

Education

Integrated M.Sc. - Applied Mathematics

Central University of Jharkhand
Ranchi
06.2019

Skills

  • Programming: Python, SQL
  • ML & DL: Scikit-Learn, XGBoost, TensorFlow, Keras, CNN
  • NLP & LLMs: Rasa, Langchain, DSPy, BERT, Mistral-7B, Claude, Phi-3, Gemini AI, Gemma-9B
  • Cloud & DevOps: AWS, Docker, FastAPI
  • Database Management: MySQL, MongoDB, Oracle
  • Data Visualization: Tableau, Matplotlib
  • Front-End: Streamlit

Certification

  • AWS Certified Cloud Practitioner
  • Tableau Desktop Professional

Projects

Yoga Pose Classification and Detection

  • Achieved 90% accuracy using MoveNet and 85% with EfficientNet for real-time pose classification.

Stock Market Data Analysis using Polygon.io API

  • Developed a Python-based data analysis tool that integrates with the Polygon.io API to retrieve and visualize historical stock market data. Utilized Pandas, NumPy, Matplotlib, and Seaborn to process and analyze trends, enabling efficient evaluation of stock performance over time.

Languages

Bengali
First Language
English
Proficient (C2)
C2
Hindi
Proficient (C2)
C2
Spanish
Beginner (A1)
A1

Timeline

AI Engineer

CLAPPIT Pvt. Ltd.
03.2024 - Current

AI Consultant

CLAPPIT Pvt. Ltd.
12.2023 - 02.2024

Data Scientist - Trainee

One Billion Ideas Pvt. Ltd.
04.2023 - 02.2024

Data Science Intern

One Billion Ideas Pvt. Ltd.
01.2023 - 03.2023

Data Science & ML Intern

Gilbert Research Center
10.2022 - 10.2022

Integrated M.Sc. - Applied Mathematics

Central University of Jharkhand
Koyana Dutta