Autonomous AI research team: write a plan.md, spawn agents, train, verify, and deliver a clean ML repo.

Real-time TUI for monitoring cloud GPU training instances

C++/Python microstructure research engine for event-driven limit order book prediction and reproducible quantitative experiments.

Selected LLM, NLP, retrieval and vision-language research-engineering projects.

Open ML systems platform for training, profiling, evaluating, and serving AI models.

Automates hermetic environments (macOS/HPC) to eliminate drift. Provisions offline RAG (Gemma 2), compiles LaTeX manuscripts, and indexes local knowledge. Unifies infrastructure, writing, and inference into a single, audit-ready artifact.

Trace-based observability and error analysis for Retrieval-Augmented Generation runs.

GitHub profile README for Nicholas Kashani Motlagh

A jekyll site for the Feast and Fast exhibition at the Fitz

Research-engineering lab for LLM post-training, behavior evaluation, regression detection, and reliability reporting.

Regime-aware portfolio using machine learning, financial time-series features, risk research engine with dynamic allocation, stress testing, factor diagnostics, reproducible research packages, and a Streamlit dashboard.

AI-powered Research Assistant using Groq, Llama 3.1, LangChain, and Tool Calling to autonomously search the web, read webpages, and generate structured research reports.

Reproducible evaluation suite for LLM behavior research: epistemic pathology, delegated introspection, and temporal consciousness diagnostics

Empirical study of neural scaling laws using transformers trained on a corpus of Python standard library modules. Investigates how model size, dataset size, and compute influence language modeling loss through Chinchilla-style scaling law fitting and small-scale transformer experiments.

Controlled mini-benchmark for context visibility, shortcut regimes, and composition in tiny causal transformers.

Adnane Arharbi

Portfolio landing page for compact AI evaluation artifacts.

A follow along of “Build a Large Language Model (From Scratch)” by Sebastian Raschka, which is explained with an insightful YouTube series “Building LLM from Scratch” by Dr. Raj Dandekar (MIT PhD).

PyTorch benchmark for CTM-style adaptive computation, sparse-retrieval failure analysis, adaptive halting, and attention-supervised recovery.

ChronosLOB: Research platform for limit order book representation learning, calibration and execution-aware validation.