Hybrid RAG API
FastAPI + Ollama exploration of retrieval, local inference, memory and modular APIs.
A clearer project archive: shipped systems first, active work second, future directions last. Each card points back to a public source where one exists.
FastAPI + Ollama exploration of retrieval, local inference, memory and modular APIs.
Applied ML project for cloud-cost prediction, with a Flask interface and prediction workflow.
C-based network scanning and analysis tooling connected to MySQL for structured scan results.
Web reconnaissance automation project built around common enumeration tooling.
Classification project for learning the practical workflow around preprocessing, training and inference.
Local-LLM-oriented security automation experiment exploring planning and tool orchestration.
Building a local-first AI developer environment around locally running models, extension workflows and GPU-aware execution.
A Python package exploring reusable AI/LLM skills, workflows, prompts and automation-oriented developer tooling.
Research and prototype direction for a backward-compatible volumetric extension of India’s ULPIN framework.
Systematic progression through neural networks, PyTorch, transformers, RAG, agents, fine-tuning, inference and deployment.
GPU-backed local inference, model discovery, developer tooling, evaluation, serving and deployment as one cohesive system.
Evaluation and monitoring patterns for local and cloud LLM systems: latency, quality, failures, cost and reproducibility.