Charles is available for hire
Hire CharlesCharles Camp
Verified Expert in Engineering
Principal AI Architect and Developer
Dijon, France
Toptal member since August 24, 2020
Bio
Charles is a principal AI architect and AWS Solutions Architect Associate with 10 years of experience delivering secure, production-grade AI for regulated industries. An ex–Credit Suisse AVP and Carnegie Mellon scholar, he specializes in private RAG, agentic workflows, and data-sovereign infrastructure. He has delivered high-impact solutions, including AML systems that cut false positives by 80% and legal platforms that draft complex briefs in under two minutes, operating on ET.
Portfolio
Safe Harbor Group, LLC
Retrieval-augmented Generation (RAG), Python, Agentic AI, Elasticsearch...
Peter Moore
Artificial Intelligence (AI), Python, Graph Databases...
Federico Lauzzana
Artificial Intelligence (AI), Large Language Models (LLMs)...
Experience
- Artificial Intelligence (AI) - 10 years
- Solution Architecture - 8 years
- Amazon Web Services (AWS) - 6 years
- Generative Artificial Intelligence (GenAI) - 4 years
- Large Language Models (LLMs) - 3 years
- Retrieval-augmented Generation (RAG) - 3 years
- Agentic AI - 3 years
- AI Agents - 2 years
Work Experience
Principal AI Evaluation & Compliance Architect
2025 - 2026
Safe Harbor Group, LLC
- Architected a production-grade, multi-agent AI compliance and evaluation framework, implementing automated quality control loops to monitor multi-step agent trajectories and model performance constraints.
- Engineered a schema-aligned document parsing and retrieval system with strict citation tracking, transforming complex regulatory parameters into auditable and defensible engineering constraints.
- Built an automated, policy-driven evaluation layer to validate agent reasoning chains, tool-use execution, and goal alignment across multi-model LLM workflows without requiring manual oversight.
- Integrated multi-model large language model (LLM) orchestration with structured prompt workflows, agent handoffs, and persistent state management.
- Deployed the full system to production on AWS with a cloud-native architecture designed for scalability.
Technologies: Retrieval-augmented Generation (RAG), Python, Agentic AI, Elasticsearch, LangChain, Legal Technology (Legaltech), LangGraph, Agentic RAG Systems, LangSmith, AI Agent Orchestration, Cloud, CTO, Early-stage Startups, New Business Ventures, Startups, Technical Leadership, Leadership
AI Architect
2025 - 2025
Peter Moore
- Extended an existing RDF4J semantic knowledge graph by building an LLM pipeline to extract causal price relationships from financial documents and inject new entity connections—enriching the graph with second-order cross-asset dependencies.
- Built AI agents using LangChain to autonomously scan financial documents, identify non-obvious causal relationships between asset price movements, and write structured connections to the semantic triplestore.
- Designed an investment opportunity detection agent that traversed the enriched knowledge graph to surface second and third-order causal chains—identifying non-obvious investment signals across asset classes.
- Built an execution agent to act on detected investment signals by placing trades on a betting platform—closing the loop from signal detection to automated execution.
Technologies: Artificial Intelligence (AI), Python, Graph Databases, Large Language Models (LLMs), Natural Language Processing (NLP), Retrieval-augmented Generation (RAG), Semantic Search, Technical Leadership
Principal AI Architect
2025 - 2025
Federico Lauzzana
- Designed and built a full-stack application from scratch—back-end, front-end, and data pipeline—ingesting real-time crypto news feeds and processing them through a sentiment classification pipeline to generate structured trading signals.
- Architected a multi-stage NLP pipeline for sentiment extraction and classification across unstructured news sources—handling noisy, high-volume financial text at scale with configurable confidence thresholds before triggering any downstream action.
- Implemented automated order execution logic based on sentiment scoring output.
- Deployed the full system to production with a scalable cloud architecture.
- Designed and implemented configurable investment strategies as structured rule sets, allowing performance tracking over time.
Technologies: Artificial Intelligence (AI), Large Language Models (LLMs), Natural Language Processing (NLP), Python, Machine Learning, TensorFlow, PyTorch, AI Agents, Reinforcement Learning, Cloud, Startups, CTO, New Business Ventures, Early-stage Startups, Technical Leadership
Principal AI Architect
2024 - 2025
Cherrington Media, LLC
- Built a full agentic API back end from scratch that takes a product description as input and orchestrates a multi-step pipeline through to finished marketing assets.
- Designed a persona identification agent that analyzes product descriptions and generates structured buyer personas—defining target audience profiles, pain points, motivations, and purchase triggers using GPT-4.
- Built a positioning agent that derives multiple distinct selling angles from the persona output.
- Implemented a marketing asset generation layer using Claude to produce campaign copy, messaging frameworks, and content variants for each selling angle.
Technologies: Artificial Intelligence (AI), Generative Artificial Intelligence (GenAI), Large Language Models (LLMs), LangChain, LangSmith, AI Agent Orchestration
Knowledge Graph Architect
2024 - 2024
Christopher Rec
- Architected ontology-driven data extraction. Integrated LlamaIndex and LangChain to ingest user-defined ontologies, enabling precise, schema-aligned entity extraction from unstructured PDFs.
- Created detailed engineering documentation, technical use cases, and product demos to showcase architecture and guide client adoption.
- Built a rigorous framework for stress testing and SDK validation, ensuring stability and accuracy of the knowledge graph pipeline under enterprise-scale data loads.
Technologies: Python, LangChain, Amazon Web Services (AWS), Knowledge Graphs, Neo4j, Vector Databases, LlamaIndex, Ontologies
LegalTech Solutions Architect
2024 - 2024
ITG AUTOMOTIVE LLC
- Automated legal contract extraction. Built a high-throughput pipeline using GPT-4 and LangChain to extract structured key terms from unstructured PDFs, accelerating document analysis.
- Standardized legal terminology. Applied hierarchical clustering to normalize diverse contract clauses into a unified schema, enabling consistent analytics and reporting.
- Deployed on AWS Lambda to ensure cost-effective, scalable document processing without infrastructure management.
Technologies: Amazon Web Services (AWS), Machine Learning Operations (MLOps), LangChain, Generative Artificial Intelligence (GenAI), Large Language Models (LLMs), Legal Technology (Legaltech), Retrieval-augmented Generation (RAG), Solution Architecture
Lead AI Architect
2023 - 2024
Huxley
- Built and deployed a secure, RAG-based AI assistant on GCP (Kubernetes/Docker) to provide personalized addiction recovery support at scale.
- Automated transcription and indexing of audio testimonials using LlamaIndex, enriching vector stores with high-value clinical context for accurate retrieval.
- Designed RL-driven optimization engines to adapt recovery pathways dynamically, improving patient engagement and outcomes.
Technologies: LangChain, LlamaIndex, Large Language Models (LLMs), Reinforcement Learning, Vector Databases, Retrieval-augmented Generation (RAG), Generative Artificial Intelligence (GenAI), Data Privacy, Solution Architecture, Startups, Early-stage Startups, Technical Leadership, CTO
Machine Learning Engineer
2022 - 2023
Fortune 500 Consumer Goods Company
- Delivered production ML pipelines at enterprise data scale for a Fortune 500 consumer goods company.
- Engineered production-ready ML code using PySpark, processing large-scale datasets within an established enterprise data platform.
- Implemented feature engineering and model workflows with scikit-learn and Pandas, ensuring functional parity and reproducibility across pipeline stages.
Technologies: Python, Machine Learning, PySpark, Scikit-learn, Pandas
Crypto Intelligence Engineer
2021 - 2022
Phragmites, Inc.
- Fine-tuned named entity recognition (NER) models to detect and classify crypto assets across high-volume, unstructured Telegram streams.
- Applied Graph Theory to map interactions and identify high-impact influencers, enabling actionable, data-driven insights.
- Developed clustering algorithms to filter near-duplicate bot messages, ensuring clean data pipelines for downstream analytics.
Technologies: Natural Language Processing (NLP), Machine Learning, Python, Data Science, Solution Architecture
AI Safety Lead
2021 - 2022
Trust & Safety Laboratory
- Developed NLP classifiers to identify harmful content and high-risk topics across high-velocity social media streams.
- Built end-to-end labeling workflows on AWS, combining automated pre-screening with human review for precise data tagging.
- Architected scalable serverless frameworks to streamline social media screening without infrastructure overhead.
Technologies: Natural Language Processing (NLP), Machine Learning, Python, Data Science, Amazon Web Services (AWS), Data Privacy, Machine Learning Operations (MLOps), Regulatory Technology (Regtech), Solution Architecture
Senior Data Scientist
2019 - 2020
Glovo
- Built and deployed scalable AWS pipelines (SageMaker, Redshift) to process high-volume transactional data for a global unicorn.
- Developed customer lifetime value (LTV) models to optimize marketing spend and acquisition across multiple markets.
- Applied linear programming to solve complex shift scheduling problems, improving warehouse efficiency.
Technologies: Data Science, Machine Learning, Python, Amazon Web Services (AWS), Machine Learning Operations (MLOps), Solution Architecture
Data Scientist AVP
2016 - 2019
Credit Suisse
- Directed a team of 8 data scientists to build an internal Adverse Media Screening NLP pilot, demonstrating strong local results and potential improvements over vendor solutions.
- Architected ML models for AML transaction monitoring that reduced false positives by 80%, achieving regulatory approval for deployment.
- Identified actionable gaps in Counterparty Risk coverage to inform risk management strategy.
- Oversaw rigorous internal validation, ensuring models met performance and compliance standards.
Technologies: Natural Language Processing (NLP), PySpark, Data Science, Machine Learning, Python, Data Privacy, Regulatory Technology (Regtech), Solution Architecture
Researcher
2016 - 2016
Carnegie Mellon University
- Developed multivariate time-series classification algorithms using EEG data to predict post-cardiac arrest survival.
- Created frameworks to detect early indicators in high-dimensional neurological datasets.
- Unsupervised Learning Insights. Applied clustering to identify latent patient subgroups for personalized treatment protocols.
Technologies: Data Science, Machine Learning
Data Scientist Intern
2015 - 2015
Capgemini
- Configured and optimized PySpark clusters on HDFS to ingest and process high-frequency industrial sensor data.
- Built ML models to detect anomalies and predict equipment failure before downtime.
- Conducted feature importance analysis to identify root causes of failures, enabling proactive maintenance.
Technologies: PySpark, Data Science, Machine Learning, Python
Project History
Privacy-preserving Federated Learning Architecture (AML)
https://github.com/SoteriaInitiative/flstandards• Architected the first open-source Federated Learning (FL) demonstrator for cross-institutional financial crime detection.
• Designed a system using Graph Neural Networks (GNN) and Differential Privacy, enabling institutions to train a shared global model without exposing client data.
• Pilot results showed significantly improved detection of complex money laundering typologies compared with isolated models.
• Designed a system using Graph Neural Networks (GNN) and Differential Privacy, enabling institutions to train a shared global model without exposing client data.
• Pilot results showed significantly improved detection of complex money laundering typologies compared with isolated models.
Agentic Legal Drafting and Retrieval System
Built a full-stack SaaS platform for federal criminal defense litigation, automating appellate brief drafting.
• Private RAG: Secure vector search pipeline using Qdrant and Elasticsearch to ensure zero data leakage.
• Agentic workflow: Specialized AI agents for fact-checking, Argument Generation, and Citation handling.
• Scalable deployment: AWS Elastic Beanstalk with Supabase for real-time case management.
• Private RAG: Secure vector search pipeline using Qdrant and Elasticsearch to ensure zero data leakage.
• Agentic workflow: Specialized AI agents for fact-checking, Argument Generation, and Citation handling.
• Scalable deployment: AWS Elastic Beanstalk with Supabase for real-time case management.
Ontology-driven Knowledge Graph Engine
• Enhanced a knowledge graph platform to support user-defined ontologies for custom entity and relationship extraction.
• Implemented a pipeline leveraging Neo4j, LlamaIndex, and LangChain for schema-driven parsing of unstructured text and PDFs.
• Led client demos and stress-testing cycles to validate improved graph traversal and extraction accuracy.
• Implemented a pipeline leveraging Neo4j, LlamaIndex, and LangChain for schema-driven parsing of unstructured text and PDFs.
• Led client demos and stress-testing cycles to validate improved graph traversal and extraction accuracy.
Education
2014 - 2016
Master's Degree in Data Science
Grenoble Institute of Technology - Grenoble, France
2011 - 2014
Bachelor's Degree in Computer Science
Grenoble Institute of Technology - Grenoble, France
Certifications
FEBRUARY 2022 - FEBRUARY 2025
AWS Solutions Architect Associate
Pearson VUE
Skills
Libraries/APIs
PySpark, TensorFlow, PyTorch, Scikit-learn, Pandas
Languages
Python
Frameworks
LlamaIndex, LangGraph
Platforms
Amazon Web Services (AWS), LangSmith
Storage
Neo4j, Elasticsearch, Graph Databases
Other
Natural Language Processing (NLP), Machine Learning, Data Science, LangChain, Large Language Models (LLMs), Machine Learning Operations (MLOps), AI Agents, Generative Artificial Intelligence (GenAI), Retrieval-augmented Generation (RAG), Data Privacy, Legal Technology (Legaltech), Ontologies, Regulatory Technology (Regtech), Solution Architecture, Agentic AI, Agentic RAG Systems, Artificial Intelligence (AI), Semantic Search, AI Agent Orchestration, Reinforcement Learning, Vector Databases, Knowledge Graphs, Federated Learning, Differential Privacy, Cloud, CTO, Early-stage Startups, New Business Ventures, Startups, Technical Leadership, Leadership
Collaboration That Works
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
1
Share your needs
Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2
Choose your talent
Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3
Start your risk-free talent trial
Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.
Top talent is in high demand.
Start hiring