About Me

Founding my way into the unknown field of AI.

My Journey

My journey started in the field of AI while I was pursuing my Master's. I developed quite a liking for the field. I started building some projects, which led me to get an internship. At that time, I learned a lot about ML algorithms and LLM architectures, fine-tuning, etc. After that, I got a chance to build agentic applications and deploy them. In my current work, I mostly build agentic applications and deploy them in edge hardware.

In my free time, I like to build projects and read research papers and blogs. I also decided to contribute to open-source projects and write blogs.

If you want to know my journey in more detail, read this blog.

Experience

Sep 2026 — Present

Artificial Intelligence Engineer

Delostyle Studio • Kolkata, West Bengal, India

  • Coming Soon...
Apr 2025 — Aug 2026

Junior Machine Learning Engineer

EEGRAB • Kolkata, West Bengal, India

  • Architected the complete AI subsystem for a fully offline educational robot (CBSE Class 1-8), leading the technical design, tech stack selection, and backend implementation logic.
  • Deployed on-premises Large Language Models using vLLM and Ollama to enable 100% offline capability, ensuring low-latency interactions via custom-hosted STT and TTS pipelines integrated with Android endpoints.
  • Engineered a Hybrid RAG system for the Teaching Mode that retrieves curriculum content from NCERT textbooks, utilizing VectorDB for semantic search and MongoDB for structural data querying.
  • Designed the Auto-Teach interactive workflow and the automated assessment module, enabling the robot to autonomously conduct scheduled lessons and generate dynamic quizzes.
  • Fine-tuned an STT model for the Bengali language, which can run on edge devices.
  • Trained and improved the accuracy of several YOLO models for factory CCTV cameras detecting PPE and further deployed them to the real-time inferencing pipeline.
  • Engineered an AI-driven automation backend using FastAPI and LangGraph, utilizing multi-tool agents to query SQL and Vector databases to generate dynamic UI charts and retrieve structured and unstructured invoice data.
Oct 2024 — Feb 2025

Data Science Intern

DaSatya • Remote

  • Designed and implemented a scalable microservices architecture using Docker and AWS EKS, developed APIs for elastic content search (S3/Local) and advanced PDF manipulation.
  • Researched and developed autonomous AI Agents using Phidata and Groq to automate Jira project management and deliver personalized news digests.
  • Optimized retrieval accuracy by experimenting with RAG Fusion and Knowledge Graphs, moving beyond basic vector search to handle complex unstructured queries.
  • Integrated MLflow for model tracking and managed the end-to-end deployment of backend services on AWS EC2.
Aug 2023 — Aug 2024

Data Science Intern

MUST Research Labs • Remote

  • Implemented a high-accuracy real-time Face Recognition system for live webcam feeds using Python and dlib.
  • Developed a RAG-based QA bot utilizing Llama-2 and LangChain, and a transactional BFSI chatbot using the Rasa framework with a FastAPI backend.
  • Fine-tuned GPT-2 models on custom datasets to automate problem-type classification and text generation for educational content.

Education

Nov 2022 — Jun 2024

M.Sc in Computer Science

University of Kalyani

Grade: 9.16

Jun 2019 — Jun 2022

B.Sc in Computer Science

Kalyani Mahavidyalaya

Grade: 8.74