# Hedi Amine Chaabani

Hedi Amine Chaabani is a remote worker on Himalayas, Lead AI Engineer at InkyLab Technologies, based in Germany.

Source: https://himalayas.app/@hediaminechaabani

```text
Lead AI Engineer at InkyLab Technologies, optimizing vLLM inference to raise throughput 150% and cut GPU cost 40%.
```

| Detail | Value |
| --- | --- |
| Availability | Looking for a job |
| Location | Germany |
| Authorized to work in | Germany |
| Salary expectations | 80k+ EUR |
| Job categories | AI Engineer, Machine Learning Engineer, Computer Vision Engineer, Software Engineer |
| Skills | LLM Inference Optimization, VLLM, Quantization, CUDA, PyTorch, RAG, Agentic AI, LLM Evaluation, LLM Monitoring, Fine Tuning, LoRAs, SFT, DPO, Distillation, Machine Learning |
| LinkedIn | https://linkedin.com/in/hedi-amine-chaabani |
| GitHub | https://github.com/HediAmineChaabani77 |

## About

```text
At InkyLab Technologies, I deployed open-weight LLMs on vLLM, raising throughput 150% and cutting GPU cost 40%. I also profiled PyTorch inference across 16 GPUs and optimized C++ and CUDA bottlenecks, lifting throughput 60% and lowering latency 35%.

At AIRE, I cut contract-chat response latency from 2.5 seconds to 800 milliseconds with vLLM serving, quantization and caching. Across my work, I’ve built RAG systems, AI agents and computer vision pipelines, including an agentic investigation platform that monitors 500 sources and 10,000 articles a day.
```

## Experience

### Lead AI Engineer (current)

InkyLab Technologies. Jul 2023 - Present

```text
Deployed open weight LLMs on vLLM, raising throughput 150% and cutting GPU cost 40%. Profiled PyTorch inference on 16 GPUs and optimized bottlenecks in C++ and CUDA, lifting throughput 60% and lowering latency 35%.
```

Skills: VLLM, Quantization, PyTorch, CUDA, C++, Python, SQL, RAG, AI-Agents, LLM, GPT, Mistral, LLaMA

### Lead AI Engineer (Full Stack)

AIRE. Jan 2026 - Aug 2026

```text
Cut LLM response latency in the contract chat from 2.5 s to 800 ms with vLLM serving, quantization and caching. Set up LLM evaluation and monitoring with an LLM judge and human in the loop review on 500+ queries, reducing hallucinations 35% and LLM cost 30%.
```

Skills: VLLM, Quantization, Caching, LLM-Evaluation, LLM-Monitoring, RAG, Python, SQL, React, TypeScript

### Senior Machine Learning Engineer

Proptexx. Aug 2024 - Aug 2025

```text
Fine tuned Stable Diffusion 1.5 and SDXL with LoRA to generate 500+ styles, cutting generation cost 70%. Deployed FastAPI and PyTorch backends serving computer vision models at 5,000+ requests a day with 99.9% uptime.
```

Skills: Stable-Diffusion, SDXL, LoRAs, fastAPI, PyTorch, Computer-Vision, CI-CD

### Senior AI Engineer

Webco Solutions Inc. Feb 2025 - May 2025

```text
Launched a RAG legal search system over 10,000+ documents that reduced search time 45%. Introduced grounded retrieval agents that cut hallucinated answers 60% across 2,000+ queries.
```

Skills: RAG, PyTorch, SQL, AI-Agents

### Machine Learning Engineer

QuickListen. Mar 2024 - May 2024

```text
Developed a Mistral LLM email assistant that reached 90% relevance across 5,000+ emails.
```

Skills: Mistral, LLM, Python, PyTorch

### Software Engineer

Integration Objects. Nov 2022 - Jul 2023

```text
Integrated machine learning models into product features, improving decision accuracy 20% for 50+ enterprise clients.
```

Skills: Python, SQL, Machine-Learning

### AI and Computer Vision Engineer

Rutilea Inc. Jan 2019 - Nov 2022

```text
Wrote C++ and CUDA code for 3D point cloud pipelines, reducing pipeline latency 60%. Shrank vision models 75% with quantization and distillation and sped up inference 250%.
```

Skills: C++, CUDA, Point-Clouds, Quantization, Distillation, PyTorch, Python, Computer-Vision

## Education

### National School of Computer Science (ENSI)

Diplôme National d'Ingénieur (Engineering Diploma), Computer Science

```text
Earned a Diplôme National d'Ingénieur in Computer Science with a focus on Data Science and Computer Vision.
```

## Tech stack

- Kubernetes
- PostgreSQL
- Django
- OpenCV
- React
- Python
- C++
- PyTorch
- MLflow
- FastAPI
- TypeScript
- Docker
- Airflow
- CUDA
- SQL
- Hugging Face
- LangChain
- vLLM
- LangGraph
- AWS
- Azure

To contact Hedi Amine, message them from this profile with a Himalayas employer account: [Hire on Himalayas](https://himalayas.app/hire).

## People also viewed

- [Salih Siddiqi](https://himalayas.app/@salihsiddiqi): Senior Machine Learning Engineer, Holi GmbH
- [ANURAG TRIVEDI](https://himalayas.app/@anuragtrivedi): ML Engineer Agentic AI, Saxony.AI
- [Romeo Alexander](https://himalayas.app/@romeoalexander): Software Engineer, Self Employed
- [Ahmed Jawad](https://himalayas.app/@ahmedjawad): Director of Engineering, Noctal.xyz
- [kabir mehta](https://himalayas.app/@kabirmehta): Machine Learning Engineer, HeidelAI
- [Ali Mirzaei](https://himalayas.app/@alimirzaei1): Co Founder & Senior AI Engineer, TutorOS
- [Rizwan Nato](https://himalayas.app/@rizwannato): Senior Machine Learning Engineer, Glanceable
- [Nandish User](https://himalayas.app/@nandishuser): Research Assistant, Otto von Guericke University / Mycrocast
- [Aleksey Sudakov](https://himalayas.app/@alekseysudakov): ML Engineer / LLM & Agentic AI Engineer, Avito
- [Haaroon Afroz Ognawala](https://himalayas.app/@haaroonafrozognawala): Research Assistant (HiWi) Deep Learning in Computer Vision, Bauhaus University
