Dr. Dipen
@Dipen
I am an AI/ML researcher with 150+ citations and 16 published research papers. I have three tier-1 publications, including Internet of Things (Elsevier), Biomedical Signal Processing and Control (Elsevier), and IEEE Access. In my research journey, I have collaborated with NASA Glenn Research Center, Cleveland Clinic, and the U.S. Department of Energy for various research projects. I am also an official reviewer and have reviewed over 100 research papers for Elsevier, IEEE Transactions, ICRA, MDPI, and other top journals and conferences. I hold a PhD from Cleveland State University with a focus on large language models (LLMs) in cybersecurity, and I also earned a master’s degree in informatics from Northeastern University.
Contributed IDE list
No IDE contribution for now
User's comments
0
articles
Figma API for AI Product Teams: Build Plugins with Cursor and Copilot for Faster Design Automation
Cursor and Copilot can compress the time to a working Figma plugin proof of concept from weeks down to hours, depending on scope and preparation. The real skill is externalizing reusable context and workflow. Clear, structured prompts finish a task in seconds; vague prompts return low-precision…Sep 24th 2026
gptq Quantization: Compare 4-Bit LLMs for Local AI Inference
GPTQ is a widely used method for compressing large models to 4-bit weights while retaining most full-precision quality. GGUF is a container file format that packages weights at various quantization levels, not a compression algorithm like GPTQ or AWQ. Full-precision large models can have very large…Sep 24th 2026
Repository-Level Code Transpilation with AlignCoder, Tests, and Retrieval
Repository-level transpilation means porting an entire codebase from one language to another while keeping cross-file dependencies, imports, and shared logic intact. AlignCoder is a retrieval-augmented framework built for repository-level code completion, and its architecture maps cleanly onto that…Sep 23rd 2026
What Is Moltbook? How AI Agents Share Tasks, Skills, and Services
The short version, before we get into it: Moltbook is described as a Reddit-style social network built only for AI agents. Agent accounts can post, comment, and vote in topic communities called submolts while humans mostly watch. Some coverage has framed it as one of the messier, more closely…Sep 22nd 2026
Reinforcement Learning Applications for LLM Agents: RFT, DPO, and SFT Compared
Three methods, three data situations. Supervised fine‑tuning (SFT) copies labeled examples. Direct preference optimization (DPO) aligns to ranked “better vs worse” pairs. Reinforcement fine‑tuning (RFT) runs a reward loop that scores outputs and pushes toward the good ones. Most reinforcement…Sep 22nd 2026
Jev is here! How to use it ?
Jev is TypeSafe AI's first System One model, as described in the vendor's materials. The product is presented as a decision engine: you hand it messy input plus a set of typed questions, and it returns structured answers with calibrated probabilities in one parallel pass. It is not designed to…Sep 21st 2026
Weighted Round Robin for LLM Inference: Route Requests by Cost and Latency
Routing LLM traffic by cost and latency stopped being optional the moment you started running interactive AI features at any real scale. Token billing and premium model prices stack up fast, and users expect answers now. Weighted round robin gives you a deterministic way to split traffic across…Sep 21st 2026
Stemming vs Lemmatization for RAG: Which Choice Improves Retrieval in LLM Apps
Stemming vs lemmatization comes down to a tradeoff between speed and linguistic precision. Stemming chops words to a crude root using fixed rules. Lemmatization maps words to their dictionary base form using vocabulary and morphology. In some retrieval settings, either normalization can improve…Sep 18th 2026
QLoRA vs LoRA for Hugging Face PEFT: Which Wins on 24GB VRAM
Watch: LoRA vs QLoRA: Fine-Tune AI Models with Just 4GB RAM! 🚀 by AI by Kiran Nagargoje For a 24GB card, QLoRA is the default for anything 7B or larger. LoRA fits comfortably only when you drop to 1-3B models. That's the decision in one sentence. Everything below is the memory, speed, and quality…Sep 17th 2026
Cursor AI vs Claude Code in 2026: LLM Agent Workflow for Real Coding Tasks
Cursor and Claude Code both start around $20/month, run on the same Claude models, and can even work side by side. What splits them is posture. Cursor assumes you're editing. Claude Code assumes you're delegating. That one difference drives everything in the table below. Which is why choosing…Sep 17th 2026
Best LLM Inference Optimization 2026: vLLM GPU Scheduling vs k0rdent AI
Quick Comparison Summary Understanding the distinction between vLLM and k0rdent AI starts with where they sit in the infrastructure stack. vLLM operates inside a single host, managing how GPU RAM stores attention states and processes concurrent requests. k0rdent AI functions at the orchestration…Sep 17th 2026
Why AI Models Need External Reviewers
External reviewers catch what internal teams miss: bias baked into training data, unverified benchmark claims, and safety gaps that only emerge under adversarial pressure. Formal independent review is still young in this discipline, and most model teams still rely on internal checks. In a bootcamp…Sep 15th 2026
Running Claude Code Agents Directly in Your Browser
Running Claude Code agents in the browser means an automated assistant can control an active web session, fill forms, navigate multi-page apps, and complete workflows you'd normally click through yourself. The barrier to entry is low: if you can run terminal commands, you can follow along. This…Sep 15th 2026
Reinforcement Learning Use Cases for LLM Alignment and Reasoning Models
Article fixed: Want a fast map of the RL techniques in this roundup? The table below sorts them by what they actually do well. We have grouped the major reinforcement learning use cases for LLM alignment and reasoning. Then we have matched each to the scenario where it earns its keep. Use it to…Sep 9th 2026
AST Explorer for LLM Code Refactors in Cursor and Copilot
Refactoring is where many bugs sneak in. That risk alone is why an AST explorer earns a spot in any LLM refactoring workflow, especially if you're still learning the craft. The core problem is old and boring: string-based edits break. Find-and-replace across a codebase falls apart the moment a…Sep 9th 2026
Designing Full-Stack Autonomous AI Agent Systems
Autonomous AI agents are software systems that reason, plan, and execute multi-step tasks. They're moving into real business operations fast. For anyone aiming at a lead AI developer role, that shift changes what "senior" means. You won't just tune models. You'll design the systems around them.…Sep 9th 2026
Build a ReAct React Web App with LangChain, OpenAI, and Google Serper
A ReAct react web app turns a static chatbot into an agent. It thinks, searches the live web, and acts on what it finds. What makes it powerful also makes it fragile. How you manage the execution flow decides how fast the app responds and how safely it handles data. Get the loop right, and you can…Sep 7th 2026
ReAct Query Builders in Python: LangChain or Custom Code for Tool-Using Agents
The LangChain-vs-custom-code choice comes down to task shape. The framework shines for open-ended, tool-rich agents with many third-party connections. Writing your own Python works better when you need full transparency, minimal dependencies, or a fixed path through the work. Both approaches do the…Sep 7th 2026
Best LLM Inference Optimization for Production Apps: vLLM, GPU Scheduling, and k0rdent AI
Did you finish a bootcamp and get a production LLM deployment? Short version: vLLM is the default engine for the best-performing LLM inference optimization, but your real cost lever is how you tune GPU scheduling granularity against latency. Miss that balance, and you either burn GPU hours or ship…Sep 6th 2026
GLM 5.2 vs Claude Fable 5 for Vibe Coding and AI App Builds
Claude Fable 5 wins nearly every benchmark. GLM-5.2 costs less on specific workloads. For a bootcamp scaling Vibe Coding projects, the GLM-5.2 vs Fable 5 call is not about raw quality. It is about matching each task to the model with the best cost per workload. Both models carry an identical…Sep 6th 2026
courses
AI Accelerator
Land an AI engineering role in as little as 90 days, without going back to school, grinding through YouTube tutorials, or needing any prior AI experience. We build your personalized roadmap, help you build a production-grade portfolio, apply for jobs on your behalf, and prep you for interviews, all the way through to a signed offer. If you don't land a role within 6 months of us starting to apply on your behalf, you get 100% of your tuition back.Jul 11th 2025
AI bootcamp 2
This advanced AI Bootcamp teaches you to design, debug, and optimize full-stack AI systems that adapt over time. You will master byte-level models, advanced decoding, and RAG architectures that integrate text, images, tables, and structured data. You will learn multi-vector indexing, late interaction, and reinforcement learning techniques like DPO, PPO, and verifier-guided feedback. Through 50+ hands-on labs using Hugging Face, DSPy, LangChain, and OpenPipe, you will graduate able to architect, deploy, and evolve enterprise-grade AI pipelines with precision and scalability.Aug 12th 2025
books
Dipen hasn't published any books