Skip to content

Hi, I’m Raufa.

AI Engineer, open to ideas and collaboration.

I work on AI models and systems, including inference optimization and quantization. I also build websites and work on UI/UX.

What I work on

Some of the things I work on with AI, web development, and UI/UX.

  • AI integration

    Connecting AI models, tools, and data sources into a workflow through MCP, harnesses, context engineering, skills, and system prompts.

    • MCP
    • Harness
    • Context engineering
    • Skill
    • System prompt
  • Agentic systems

    Building agents and automation to carry out multi-step tasks and reduce repetitive work, using Hermes or a Python harness.

    • Automation
    • Hermes
    • Custom Python harness
  • RAG and NLP

    Document search and question answering. Combining BM25, embeddings, and rerankers to retrieve relevant context before generating an answer.

    • RAG
    • BM25
    • Embedding
    • Reranker
  • Speech

    ASR and TTS, plus speech restoration for recordings with echo or background noise.

    • ASR
    • TTS
    • Speech restoration
    • Dereverb
    • Denoise
  • Computer vision

    Object detection with YOLO, pose estimation through joints and keypoints, and image classification from images or video.

    • YOLO
    • Object detection
    • Pose estimation
    • Image classification
  • Fine-tuning and training

    Adapting models to a dataset or domain through LoRA, PEFT, partial or full fine-tuning, CPT for domain adaptation, or training from scratch.

    • LoRA
    • PEFT
    • Partial fine-tuning
    • Full fine-tuning
    • CPT
    • Training from scratch
  • Quantization and inference

    Optimizing inference through quantization at different precisions and kernel optimization to reduce a model’s compute requirements. My explorations also include ternary quantization.

    • Inference optimization
    • Kernel
    • Quantization
    • Ternary quantization
  • Dataset preparation

    Preparing training data: collecting, cleaning and formatting it, since the data decides how well the model learns.

    • Dataset preparation
    • Data cleaning
  • AI problem solving

    Finding ways to use AI to solve everyday problems or make daily tasks easier, while following new tools and understanding their limitations.

    • AI literacy
    • Problem solving
    • AI tools
  • Web and UI/UX

    Building websites and interfaces, thinking through how people use a product as well as how it looks and works.

    • Web development
    • UI/UX
    • React
    • SvelteKit

Languages and technologies

  • Python
  • Go
  • JavaScript
  • TypeScript
  • React
  • SvelteKit
  • C++

How I find a solution

  1. Brainstorm & research

    Understand the problem, discuss ideas, and investigate possible approaches.

  2. Plan

    Work out the next steps from what we have discussed and learned.

  3. Build the solution

    Put the agreed plan into practice.

A little about how I work

What are you focused on?

My focus is AI engineering: AI integration, agents, RAG, ASR, TTS, computer vision, training, and dataset preparation. I also explore inference optimization and quantization.

Do I need a project plan before getting in touch?

No. We can start with an idea, a technical question, or an everyday problem to work through. A plan can follow once the context is clearer.

Do you also work on web and UI/UX?

Yes. Alongside AI, I build websites and interfaces and understand the basics of UI/UX and system architecture.

Do you work independently or with a team?

Both. I can work independently or collaborate with a team, depending on the needs and the role we discuss together.

How do you start a new project?

I begin with brainstorming and research to understand the problem, then plan and work on the solution.

Anything else you would like to discuss? Send a message.

Have an idea or something to think through?
Let’s talk.

Open to technical discussions, collaboration, and projects. We can start by getting to know the idea.