Adel Basli — AI/ML Leader

Profile

Strategic AI Leader with 11+ years of experience specializing in scaling Large Language Models (LLMs) and autonomous agentic workflows. Proven track record directing international teams (25+ members) to deploy production-grade AI systems.

Expert in Foundation Model fine-tuning (SFT/RLHF/DPO), RAG architecture, and MLOps at scale. Deeply focused on bridging the gap between research and high-impact business automation.

25+ Engineers

Led across 3 cross-functional teams globally

$12MM+ Impact

Revenue generated via AutoML platform

82% Automation

Manual labeling replaced by agentic workflows

60,000+ Models

Deployed to production globally

Technical Skills

Generative AI & LLMs

LLM Fine-tuning (RLHF/DPO/SFT) LoRA / QLoRA RAG Architecture Agentic Workflows LangGraph MCP Prompt Engineering Quantization (GGUF/FP8) Model Distillation LLM-as-a-Judge

Foundation Models & Frameworks

Gemini 1.5 Pro Llama 3.x Claude 3.5 GPT-4 PyTorch JAX Hugging Face Transformers PyTorch FSDP DeepSpeed vLLM

MLOps & Infrastructure

Vertex AI AWS SageMaker MLflow Kubernetes (GKE) Docker CI/CD Pinecone FAISS Elasticsearch Apache Spark

Programming & Data

Python (Expert) SQL BigQuery Snowflake Apache Spark Hive TensorFlow Scikit-learn Pandas NumPy

AI Safety & Evaluation

Hallucination Detection Bias Auditing Differential Privacy Red Teaming Responsible AI Model Evaluation

System Design & Leadership

Distributed Systems Microservices Agile / Scrum Technical Roadmapping Cross-functional Leadership FinOps

Professional Experience

Team Lead Data Science & AI — General Manager

NielsenIQ (Label Insight) · Chicago, USA

06/2025 — Present

Directing the global AI function (3 cross-functional teams, 25+ associates) overseeing the lifecycle of autonomous data capture and labeling systems.
Responsible AI Evaluation: Architected an automated "LLM-as-a-Judge" framework for hallucination detection and bias auditing, ensuring 99.9% safety compliance across production agents.
Agentic Orchestration: Designed multi-agent workflows using LangGraph and MCP to automate 82% of manual labeling, improving processing speed by 35%.
Strategic Roadmap: Formulated a 3-year AI transformation plan resulting in a 65% improvement in automated data enrichment efficiency.

Team Lead Data Science & AI — Functional Manager

NielsenIQ (Label Insight) · Paris, France

10/2023 — 05/2025

Data-Centric AI: Developed a synthetic data generation engine using LoRA-finetuned models to bootstrap training for low-resource categories, increasing model coverage by 45%.
Multimodal Search: Integrated a multimodal vector database (Pinecone) with CLIP-based embeddings, reducing search latency by 60% for product discovery.
Generative AI & Search: Developed outlier correction models using RAG and Elasticsearch, reducing manual oversight by 27%.
Managed FinOps for cost-effective data science solutions, reducing operational costs by 20% while maintaining performance.

Lead Data Scientist

NielsenIQ (Label Insight) · Chicago, USA

01/2023 — 10/2023

Built an LLM-based classifier saving $3MM/year in data transformation costs by replacing legacy heuristics.
Edge Optimization: Led quantization (GGUF/FP8) and distillation of local models for on-premise deployment, ensuring high performance under strict data privacy constraints.
Deployed scalable real-time data system on AWS, reducing processing time by 40%.

Senior Machine Learning Engineer — R&D Lead

Nielsen IQ · Chicago, USA

01/2021 — 12/2022

Distributed Training: Optimized model training pipelines using PyTorch FSDP, reducing training time for 7B+ parameter models from weeks to days.
Directed R&D on Transformer architectures, yielding a 5% gain in prediction accuracy.
Built stacking ensembles automating predictive tool deployment for 9+ global server environments.

Data Operations Transformation Lead

The Nielsen Company · Chicago, USA

01/2019 — 12/2020

Developed an AutoML platform generating $12MM in incremental revenue and cost savings.
Privacy Engineering: Implemented Differential Privacy (DP) protocols for sensitive consumer datasets, maintaining model utility while meeting global compliance standards.
Led the deployment of a portfolio containing 60,000+ production models globally.

Senior Data Scientist

The Nielsen Company · Paris, France

01/2016 — 12/2018

Leveraged MapReduce and Big Data ETL to reduce M&A integration timelines by 6 months.
Deployed an unsupervised item coding system using fuzzy matching, generating $400,000 in additional revenue.

Data Scientist

Air Liquide · Newark, USA

03/2014 — 12/2015

Developed predictive tools for gas demand forecasting, reducing inventory costs by 15%.
Created a process control tool using DTW clustering, saving $500,000 annually.

Selected Open Source & Community

Open Food Facts

Lead AI Contributor

Vision-Language Pipeline: Architected a VLM-based OCR system to extract and validate nutritional facts from 3M+ product images, improving ingestion speed by 300%.
Nutri-Score Automation: Built a deep learning classifier to predict Nutri-Score categories from ingredient lists.

View Project