Taher Nouira

About Me

I'm a Software Engineering student at INSAT, focused on building reliable AI systems across LLMs, GenAI, and Computer Vision.

I care about understanding models from the inside out: implementing transformers from scratch, designing retrieval-augmented generation pipelines, and hardening LLMs with safe, production-grade tooling. Most of what I build sits at the intersection of research, engineering, and security.

Current Focus

LLMs · RAG · Multimodal · LLM Security

Strengths

Systems thinking, end-to-end pipelines, hands-on implementations

Open To

AI internships · Research-driven engineering roles

Focus Areas

AI & GenAI

Large language models, retrieval-augmented generation, prompt engineering, and fine-tuning — with a focus on understanding failure modes and alignment with real data.

Computer Vision

Visual document understanding, multimodal models, and practical computer vision systems that support classification, detection, and decision pipelines.

ML Systems

From prototype to production: APIs, orchestration, observability, and scalable deployment of AI services.

Research to Code

Turning papers into running systems: re-implementing architectures, evaluating tradeoffs, and shipping minimal, understandable versions.

Experience

AI / NLP Engineer

Dydon AI · Herrliberg, Switzerland

Apr 2025 – Oct 2025
  • Optimized a RAG pipeline for financial document understanding, improving retrieval precision by 15% and response coherence.
  • Integrated FlashRank and lightweight rerankers, achieving a 2.3× speedup while preserving answer quality.
  • Built a prompt sanitization & LLM security layer (regex, escaping, rewriting) deployed via Docker & FastAPI.
  • Developed a visual document classifier using MobileNetV2 + OpenCV EPC pipeline with 100% accuracy across FR / DE / AT.

Software Developer Intern

BEDI · Monastir, Tunisia

Jun 2024 – Sep 2024
  • Built an attendance system with PyQt and ZKTeco biometric SDK for real-time authentication.
  • Designed fault-tolerant microservices using RabbitMQ for asynchronous communication.

Skills & Technologies

Languages

PythonTypeScriptJavaScriptSQLC++Bash

ML & AI Frameworks

PyTorchTransformersLangChainOpenAI APIAnthropic APIChromaDBFAISSCUDAscikit-learnNumPyPandas

Development & Tools

GitDockerFastAPIReactAstroNode.jsPostgreSQLMongoDBLinuxVS Code

Cloud & MLOps

AWSGoogle CloudCloudflareVercelJupyterWeights & Biases

Let's Build Something Useful

I’m interested in roles where I can design, implement, and stress-test AI systems end-to-end — from model internals to infrastructure. If that sounds aligned, feel free to reach out.