Hi, I’m Ameni Khabthani
I'm a
AI Engineer and Data Scientist specializing in speech technologies, NLP, and end-to-end machine learning systems. I work with startups and teams to build and deploy reliable, production-ready AI solutions.
Ameni Khabthani
AI & Voice Engineer | Freelancer12+
Projects
6
Clients
1+
Years Experience
Looking to Build Practical AI Solutions?
I build end-to-end speech and language AI systems, including automatic speech recognition, text-to-speech, dubbing, translation, and large language models. My work focuses on leveraging state-of-the-art approaches to deliver reliable, high-quality solutions.
I emphasize clean data pipelines, efficient training and finetuning strategies, and scalable inference workflows designed for real-world deployment across multiple languages and accents.
I deliver end-to-end AI solutions in speech technologies and NLP, covering dataset preparation, model finetuning (LLMs, ASR, TTS), and deployment of production-ready systems.
Transform Your Vision Into AI-Powered Solutions
Whether it’s speech recognition, text-to-speech, or large language model finetuning, I help turn ideas into fully functional AI systems.
View case studies
A summary of my professional experience, freelance projects, and academic background.
Freelance Projects
Selected freelance projects focused on speech processing, NLP systems, and end-to-end AI pipelines.
Medical ASR & Text-to-Text Translation System
Lexi
[In Progress]
Working on a medical speech recognition and text-to-text translation system, aimed at improving multilingual clinical communication.
Seamless Dubbing Automation
Saas startup (NDA)
Context
Built a fully automated multilingual dubbing system to replace a costly and inflexible third-party service.
The solution transformed raw audio and video content into synchronized, natural-sounding dubbed outputs across multiple languages.
Realisation
- Supported both audio- and video-based inputs for end-to-end dubbing.
- Implemented speaker diarization and ASR using Whisper Large v3.
- Performed automatic translation and multilingual speech synthesis using a hybrid XTTS + Orpheus TTS approach.
- Applied audio alignment and merging to produce synchronized, natural-sounding dubbed output.
- Deployed the full pipeline as a Flask API on Azure App Service.
Impact
Enabled scalable, high-quality multilingual dubbing with improved voice consistency, accurate synchronization, and full control over the end-to-end process.
AI-based Astrology API
Independent Client
Context
Designed an AI-powered astrology API that delivers personalized, context-aware answers to horary astrology questions. By combining traditional astrological logic with modern NLP and large language models, the system produces coherent and meaningful interpretations.
Realisation
- Designed a backend API delivering personalized astrology responses.
- Integrated an LLM (GPT-4o) with prompt engineering and intent classification to generate context-aware, user-specific insights.
- Combined traditional astrology rules with AI-driven reasoning for more coherent and meaningful outputs.
- Implemented database management using Supabase to store user profiles, query histories, and custom configurations.
- Designed the system for extensibility, allowing dynamic prompts and configurable response logic.
Impact
Provided scalable, personalized astrology experiences with consistent, context-aware responses, improving user engagement and allowing flexible customization for different user needs.
Creating a Custom TTS System
Independent Clients
Context
Designed and fine-tuned a custom text-to-speech system capable of voice cloning and speaker adaptation for natural, expressive speech generation, positionned for multiple clients across different languages including Tunisian, Maghrebian, Hindi, and Hinglish.
Realisation
- Prepared and curated multilingual datasets in the required format for fine-tuning for each client.
- Applied audio preprocessing, normalization, and segmentation to ensure high-quality training data.
- Conducted full model fine-tuning to capture unique speaker characteristics and multilingual nuances.
- Implemented inference in both batch and streaming modes based on client requirements.
- Delivered a flexible TTS system adaptable to different languages, accents, and deployment preferences.
Impact
Produced highly natural, expressive, and multilingual speech output for multiple clients, offering them a flexible and scalable solution for their audio content needs.
Academic Background
My academic background, complemented by professional certifications.
Master’s in Data Science, AI, and HealthTech
Pristini School of AI
Bachelor’s Degree in Electrical Engineering Specializing in Automation and Industrial Computing.
Iset Mahdia
Professional Certifications & Badges
Explore my AI and Voice Technology demos, including Speech Recognition and Text-to-Speech projects.
- Recent achievements
Like what you see?
Check out my demos and see the power of AI-driven speech technology in action.
What Clients Say
Abhishek Kaul
Entrepreneur, NextKid Private limited
I had the pleasure of working with Ameni Khabthani as a client, and I can confidently say she exceeded my expectations in every aspect.
She freelanced for my company NextKid Private Limited, where she developed a high-quality child voice Text-to-Speech (TTS) system by finetuning Coqui TTS. The results were outstanding — the voice quality was natural, expressive, and perfectly suited for our use case. Her technical expertise in voice AI and finetuning models is truly impressive.
Ameni delivered everything on time, paid great attention to detail, and executed the project exactly as discussed. Beyond her strong technical skills, her professionalism, communication, and positive nature made the collaboration smooth and enjoyable. She was responsive, open to feedback, and always focused on delivering the best possible outcome.
I would highly recommend Ameni to anyone looking for a skilled, reliable, and professional AI/voice engineer. I'd be happy to work with her again in the future.
Feel free to reach out for AI projects, speech technology collaborations, freelance opportunities, or technical consultations. I'm always excited to explore innovative ideas.