| π Education | πΌ Research Experience | π¬ Publications | π Location |
|---|---|---|---|
| B.Tech CSE (AIML) The Neotia University (CGPA: 9.42) |
AI/ML Research Intern NIT Durgapur |
4 Papers IEEE & Springer |
West Bengal, India |
Hey there! π I'm Arpan Pramanik, an AI/ML Engineer & Full-Stack Architect passionate about bridging cutting-edge artificial intelligence with resilient, production-grade cloud systems. I specialize in designing autonomous agentic workflows, low-latency RAG architectures, physics-grounded agronomy engines, and observable multi-portal enterprise SaaS.
- π§ Production Generative AI & Agents: Architecting resilient RAG systems, multi-LLM failover routing, dense vector stores (
pgvector,ChromaDB,FAISS), and anti-hallucination guardrails. - π¬ Scientific Rigor & Explainable AI (XAI): Designing custom multi-headed CNNs and transfer learning models with Grad-CAM interpretability for real-world automated inspection.
- β‘ High-Performance Full-Stack SaaS: Building reactive microservices with Next.js, React, FastAPI, Node.js & Express with real-time SSE telemetry and role-based access control.
- π Enterprise Observability & Cloud: Instrumenting production distributed tracing, structured logging, and metrics using OpenTelemetry, Prometheus, Pino, Docker & AWS.
| Category | Technologies & Tools |
|---|---|
| Languages |        |
| AI, ML & RAG |          |
| Backend & APIs |       |
| Frontend & UI |      |
| Databases & Stores |      |
| Cloud, DevOps & Observability |       |
π 1. PaperLens AI β RAG-Powered Research Paper Assistant
AI-Driven Academic Research Platform with Persistent Vector Search & Multi-LLM Failover Routing
- β‘ Production Vector Storage: Persistent semantic vector search with Supabase
pgvector& FAISS paired with generator-based PyMuPDF extraction. - π― Citation Validation Pipeline: 4-stage automated validation (
DOI β Exact β Title β Loose) via Semantic Scholar API with real-time Server-Sent Events (SSE) streaming. - π§ Resilient Multi-LLM Routing: Automatic failover powering 6+ AI research modules (Contextual Q&A, Experiment Planner, Research Gap Detection, and Problem Generator).
π§ 2. JalDrishti AI (ΰ€ΰ€²ΰ€¦ΰ₯ΰ€·ΰ₯ΰ€ΰ€Ώ) β Climate-Smart Precision Agronomy Platform
Sensorless FAO-56 Hydrology & Precision Irrigation Advisory Platform Engineered for Indian Smallholders β’ GitHub Repository β’ Documentation
- π§ FAO-56 Hydrology & Rain Hold Engine: Daily Penman-Monteith ETβ calculations, continuous 4-stage crop phenology curves (
$K_c$ ), persistent root-zone depletion balance, and automated rain-hold cost guards saving βΉ150ββΉ500/run in pump wastage. - π Pathogen Alerting & Financial ROI: Daily microclimate pathogen proliferation models for 5 staple crops (Paddy, Potato, Wheat, Mustard, Maize) alongside idempotent accounting tracking Liters Saved, Pump Hours Avoided, and avoided COβ.
- π€ JalSathi AI Voice Assistant: 100% Groq-native multilingual RAG (Bengali, Hindi, English) with sub-150ms query translation, ChromaDB vector search over ICAR guides, 3-tier offline fallback, and strict 45+ agrochemical safety guardrails.
π 3. CampusSphere β Observable Student Management Platform
Enterprise Multi-Portal Academic SaaS with Comprehensive OpenTelemetry Distributed Tracing
- π 3-Portal Architecture: Dedicated Admin, Faculty, and Student portals exposing 25+ secure REST API endpoints engineered for institutional compliance.
- π‘οΈ Production Observability Telemetry: Hardened JWT authentication, strict CORS, rate-limiting, Pino structured logging, and end-to-end OpenTelemetry distributed tracing paired with Prometheus monitoring.
- β‘ Scalable Cloud Stack: Server-rendered frontend interface built with Next.js & Tailwind CSS backed by high-throughput Express.js services and relational PostgreSQL data models.
- π₯ Authors: Shibdas Dutta, Subhrendu Guha Neogi, Diya Chanda, Arpan Pramanik, ΓzgΓΌn Girgin, Enes Ladin ΓncΓΌl
- π DOI: 10.1109/ICRITO66076.2025.11241706
- π‘ Summary: Multi-task deep learning framework for fruit classification and quality assessment using multi-headed CNN. 98% classification & 99% quality detection accuracy with Grad-CAM interpretability.
π Paper 2: CropSense: Explainable Deep Learning Framework for Accurate Quality Detection in Solanaceous Crops
- π₯ Authors: Shibdas Dutta, Subhrendu Guha Neogi, Shiladitya Chowdhury, Vikrant Chole, Arpan Pramanik, Diya Chanda
- π DOI: 10.1109/ICRITO66076.2025.11241535
- π‘ Summary: Lightweight multi-headed CNN for potato and tomato quality classification. 99.9% crop classification & 98.5% quality detection accuracy with Grad-CAM, deployed via Streamlit.
- π₯ Authors: Shibdas Dutta, Barshan Adhikari, Arpan Pramanik, Diya Chanda
- π DOI: 10.1109/COMPUTINGCON64838.2025.11376762
- π‘ Summary: Hybrid CNN-ViT model for simultaneous crop classification and quality assessment. Reduces parameters by 30%+. 98.45% potato & 97.49% tomato classification accuracy with 98.5% quality assessment.
π Paper 4: Hyperspectral Fruit and Vegetable Classification Using Convolutional Neural Networks with EfficientNetB3
- π₯ Authors: Shibdas Dutta, Subhrendu Guha Neogi, Arpan Pramanik, Diya Chanda, ΓzgΓΌn Girgin, Enes Ladin ΓncΓΌl
- π DOI: 10.1007/978-3-032-21901-5_35
- π‘ Summary: Transfer learning framework with EfficientNetB3 across 4,320 hyperspectral images over 36 classes. 99.29% training & 97.21% test accuracy for intelligent sorting, deployed via Streamlit & OpenCV.
| Domain | Skills & Architectures |
|---|---|
| Machine Learning & Deep Learning | Supervised & Unsupervised Learning, Feature Engineering, Multi-Headed CNNs, EfficientNet, Transfer Learning |
| Computer Vision & Explainable AI | OpenCV, Image Classification & Grading, Grad-CAM Attention Heatmaps, Streamlit Deployments |
| LLMs, RAG & Autonomous Agents | LangChain, Groq, Ollama, Vector Search (pgvector, ChromaDB, FAISS), Multi-LLM Routing, STT/TTS |
| Full-Stack SaaS & Web | Next.js, React, Tailwind CSS, FastAPI, Express.js, Node.js, REST APIs, SSE Event Streaming |
| Databases & Stores | Supabase, PostgreSQL, MongoDB, Redis, Pinecone, ChromaDB, MySQL |
| Cloud, DevOps & Observability | AWS (EC2/S3/Lambda), Docker, OpenTelemetry Distributed Tracing, Prometheus, Pino, GitHub Actions |
πΌ Open to: High-Impact AI/ML Engineering Roles, Production SaaS Development, Research Collaborations
β‘ "Code is like humor. When you have to explain it, it's bad." β Cory House