← Back to Portfolio
Download Original PDF

Debapriya Das

Pune, Maharashtra, India | debapriyadas73@gmail.com | +91 6295747430 | dev.to/debapriyadas
linkedin.com/in/debapriya-das-510728220 | github.com/Debapriya-source

Summary

AI/ML enthusiast with industrial experience in generative AI, backend development, DevOps, and agentic AI development for 2+ years. Passionate about building scalable systems, exploring intelligent solutions, and working with impactful technologies.

Experience

Senior AI Engineer (Full-Time), Kvell Dynamics, Remote, South Africa
Oct 2024 – Present
RMB — Rand Merchant Bank (Project Lead)
  • Leading the development and DevOps team of 10+ members.
  • Creating GCP cloud architectures for efficient CI/CD pipelines.
Fungi — Smart Utility Management Platform (Project Lead)
  • Architected and developed a large-scale FastAPI backend with 60+ REST API routers for managing sites, meters, customers, billing, tariffs, and anomaly detection.
  • Designed PostgreSQL data layer using SQLModel/SQLAlchemy with Alembic migrations and optimized connection pooling.
  • Built billing rule execution engine, tariff calculation system (block rate, time-of-use, demand charge), and payment processing integration.
  • Implemented Redis caching, RabbitMQ background job processing, and event-driven architecture using FastStream.
  • Built headend ingestion layer for 4 meter systems (M2M, Landis & Gyr, Sigfox, Scigrow) with async batch processing, delta calculation, and device-type routing (electricity, water, gas).
  • Led data migration from legacy system to new platform — PostgreSQL-to-PostgreSQL normalization and PostgreSQL-to-InfluxDB time-series migration with validation pipelines.
  • Developed AWS Lambda functions, Cognito auth flows, and SES email services using boto3; implemented JWT + Casbin RBAC, Docker containerization, and CI/CD pipelines.
Murphy's Law AI — Legal AI (early phase, ~1 month)
  • Built conversational AI interface using Chainlit, integrated LLM APIs, and developed file upload feature.
  • Implemented RAG pipeline and ETL using Weaviate vector database and Langchain.
Hackathon Intern, ANNAM.AI, Rupnagar, Punjab, India
May 2025 – Jul 2025
  • Participated in a hackathon-based internship organized by ANNAM.AI, an AI CoE of the Ministry of Education, Govt. of India.
  • Developed AI solutions for real-time challenges in Indian agriculture under mentorship from domain experts.
Full Stack Engineer (Freelance), Nearhouse, Santiniketan, India
Sept 2022 – Jun 2024
  • Developed backend APIs with Express.js from scratch. Live: nearhouse.in
  • Database: MongoDB. Implemented CI/CD pipeline for automated deployment from GitHub to AWS EC2 using GitHub Actions.
Backend Engineer (Freelance), Lazer, Remote, India
Dec 2021 – Dec 2022
  • Created REST APIs using Express.js, used MongoDB for data storage, managed AWS CI/CD pipeline.

Education

IIT Ropar, Major in Artificial Intelligence
2024 – 2025
  • Coursework: Python, Artificial Intelligence, Prompt Engineering, NumPy, Pandas
Visva-Bharati University, M.Sc. in Computer Science (CGPA: 7.75/10.0)
2022 – 2024
  • Coursework: Advanced Machine Learning, Image Processing, Data Mining, Distributed Systems, Parallel Algorithms, Approximation Algorithms, Advanced Software Engineering, Quantum Computation
Visva-Bharati University, B.Sc. in Computer Science (CGPA: 7.85/10.0)
2019 – 2022
  • Coursework: C Programming, Computer Architecture, Data Structures & Algorithms, Operating Systems, DBMS, Theory of Computation, Numerical Methods, Artificial Intelligence, Soft Computing, Computer Networks, Linux Programming

Skills

Languages: Python, JavaScript, Node.js, C, C++, CSS, HTML, Java, Dart

Technologies: AWS, MySQL, MongoDB, PostgreSQL, REST API, OpenCV, NumPy, Scikit-Learn, Matplotlib, NLTK, GitHub, Pandas, DevOps, Docker, Langchain, Redis, InfluxDB, Weaviate

Frameworks: PyTorch, TensorFlow, React.js, Express.js, FastAPI, Streamlit, Chainlit, FastStream

Research

Generation of de novo Molecular Structures using Graph-based Variational Autoencoder
DCSS, Siksha Bhavana, VBU, India
Mar 2023 – Jun 2024

Developed a generative AI model from scratch for producing novel molecular structures from graph inputs using GNNs and VAEs.

  • Developed a novel Graph-VAE architecture with edge-informed graph convolutions.
  • Introduced a quadratic O(n²) complexity reconstruction loss, replacing quartic O(n⁴) complexity.
  • Preprocessed and fine-tuned on ESOL & TOX21 datasets. Enhanced chemical validity from 32% to nearly 98%.
  • Tools: Python, PyTorch, DGL, DGLlife, RDKit, NumPy, Pandas

Key Projects

Nyaya-GPT – A RAG+ReAct based framework (NLP)
Sept 2024
  • RAG framework on Indian Constitution and BNS powered by a ReAct loop. (83+ ⭐ Stars on GitHub)
  • Tools: Python, Groq, Streamlit, Llama 3.1-70b, Langchain
TaxBuddy / The Tech Buddy – Customized AI Chatbot (NLP)
Sept 2024
  • Agentic RAG chatbot for Indian TAX related Q&As using DeepSeek R1 and Llama 3.
  • Tools: Python, Groq, Streamlit, Llama 3.1
Custom Zebra Fish Detection – YOLOv5 (Computer Vision)
Jun 2023
  • Extracted frames using OpenCV, labeled dataset, fine-tuned YOLOv5 model.

Publications & Articles

A study on erratic mental status in a young human population in respect to mate choice (Dec 2023)
sciforum.net/paper/view/16541
Exploring Word Embedding Techniques Based on Count or Frequency: A Practical Guide (Jul 2024)
Published on LinkedIn. Covers One-Hot Encoding, Bag of Words, and TF-IDF techniques for NLP.
Exploring Text Preprocessing Techniques in Natural Language Processing (Jul 2024)
Published on LinkedIn. Covers foundational text preprocessing techniques essential for NLP tasks.

Certifications

Introduction to Machine Learning (2021) | Duke University, USA
Covered logistic regression, MLPs, CNNs, and NLP with hands-on PyTorch exercises.
Applications of Python in Big Data, NLP, and Social Network Analysis (2020) | IEEE Kolkata Chapter
Hands-on workshop on Python applications in Big Data, NLP, and SNA.