Charles Camp
Verified Expert in Engineering
Machine Learning Developer
Charles is certified in artificial intelligence and data science. He is highly skilled at producing high-performing models and making them easy to use. He easily adapts to all kinds of environments and has already worked for banks, startups, IT firms, and laboratories. Charles' fields of expertise are natural language processing and time series analysis.
Portfolio
Experience
Availability
Preferred Environment
Python, Amazon Web Services (AWS), Natural Language Processing (NLP), Time Series Analysis, Transformers, Reinforcement Learning
The most amazing...
...model I've built can identify people with a connection to financial crimes.
Work Experience
AI Engineer
ITG AUTOMOTIVE LLC
- Extracted key details from contracts using ChatGPT, LangChain and Pydantic.
- Benchmarked a hierarchical clustering algorithm to group extracted key terms based on their content and organize them.
- Integrated the solution and deployed it on AWS using Lambdas.
AI LLM Expert
Huxley
- Built a chatbot trained on deaddiction program literature to help people get sober for different types of addictions, like alcohol, overeating, etc.
- Built the API of the chatbot and deployed it on GCP in a scalable manner using Docker and Kubernetes.
- Transcribed audio testimonials to enrich the quantity of data available for the chatbot's recommendations.
- Used llama-index to improve the indexing of the vector store (embeddings of the documents) and get better-quality answers from the chatbot.
- Built a model to extract quotes and stories from documents to use them as testimonial summaries.
- Trained a recommender system to suggest inspiring quotes to the users based on their profiles. This included experience in the program, past likes/dislikes, inventories, and app activity.
- Produced documentation and had a live session with the founder to explain the technicalities of the solutions.
- Trained a reinforcement learning model to shorten the path to sobriety. Used this model to choose what actions the sponsor should take to make the user evolve as fast and reliably as possible in the program and remain sober.
AI Developer
Non-Fungible Films, Inc.
- Fine-tuned stable diffusion models to be usable with the company's fictive characters.
- Deployed a Discord bot similar to Midjourney but using the custom stable diffusion model.
- Integrated the model with a stable diffusion UI to enable inpainting, image to image, and other applications.
ML Engineer
Global CPG Company
- Created a pipeline to automate look-alike audience computations leveraging internal consumer behavior data.
- Compared models to achieve the highest performances and hyperparameters tuning.
- Created custom PySpark and scikit-learn estimators to integrate with PySpark and scikit-learn pipelines, respectively.
ML and NLP Engineer
Phragmites, Inc.
- Set up an EC2 server, analyzed Telegram messages stored on a Postgres DB, and classified them as relevant or not for a given crypto-related project.
- Built a bot message-detection model using near-duplicates clustering approaches.
- Quantified the influence of Telegram users in crypto-centered conversations using graph theory.
- Trained a NER model to detect crypto project names in Telegram messages.
Senior Data Scientist
Trust & Safety Laboratory
- Trained machine learning models to find controversial topics in tweets. Controversial topics were defined as potentially containing harmful misinformation.
- Trained ML models to detect false claims and misinformation in tweets.
- Built a pipeline to collect human loop reviews (AWS), automated the labeling of potentially misleading tweets, and performed website scraping.
- Developed a serverless framework to automatize social media screening tasks.
Python Developer | AI
Click Factura SA de CV
- Transcribed and summarized Spanish audio meetings: fine-tuned text-to-speech models ( DeepSpeech, NeMo, and Wav2Vec) and used text summary and diarisation models.
- Trained an OCR model to extract the information on Mexican expense tickets.
- Integrated the models into the existing Django application by creating APIs.
- Deployed the models using Docker containers and Flask.
Machine Learning Expert | Digital Advertisement
Primal Analytics
- Deployed Lambda to detect anomalies in Google ads statistics automatically.
- Compared various state-of-the-art ML models for time-series anomaly detection.
- Set up the AWS account for data storage and Lambda execution.
Senior Data Scientist
Glovo
- Designed, implemented, and deployed a customer lifetime value model. This was deployed on an EC2 instance using Luigi and scheduled with Jenkins.
- Used linear programming to optimize pickers' time shifts.
- Built an end-to-end pipeline to decide whether or not to show a product in the app based on its probability of being available in the store to improve the customer experience. The model was trained on SageMaker and then deployed on an EC2 instance.
Data Scientist
Credit Suisse
- Designed and deployed machine learning models to detect money laundering using transaction data.
- Led the Negative News screening project to automatically screen news data and find associations with financial crimes to enrich the risk scoring model.
- Used NLP to measure the impact of news data on the sales of financial products.
- Organized data sourcing and mapping of various transaction and KYC data sources on a big data platform. Also handled the design and implementation of a data model for the transaction and KYC data to facilitate transaction monitoring.
Research Scholar
Carnegie Mellon University
- Designed and implemented a model to predict the survival of patients after a cardiac arrest using their brain activity data (multivariate time series).
- Built an evaluation to give a better score to models predicting the survival earlier.
- Clustered patients to identify common characteristics and deduce specific preventive actions to increase their survival rate.
Data Scientist Intern
Capgemini
- Set up a Spark cluster to read sensor data from HDFS and preprocess it.
- Built a scalable supervised model to detect manufacturing breakdowns using multivariate time series data (sensor data).
- Fine-tuned and validated the model. Identified main features leading to breakdowns.
Experience
Recommender System
In the first step, we use non-negative matrix factorization (NMF) to find two matrices W and H of respective sizes (number of users, K) and (K, number of movies) that minimize the difference between V and WH where K is a small value (< 10). That means we look for W and H such as WH is close to V.
Afterward, we use W and H to cluster the users and can now recommend movies that will be liked by their assigned cluster.
Face and Image Recognition
Skills
Languages
Python, SQL, R
Libraries/APIs
Scikit-learn, Pandas, PySpark, SpaCy, Natural Language Toolkit (NLTK), XGBoost, TensorFlow, Luigi, OpenCV, Node.js
Paradigms
Data Science, Anomaly Detection, Linear Programming, Test Automation
Other
Time Series Analysis, Natural Language Processing (NLP), Machine Learning, Artificial Intelligence (AI), Communication, GPT, Generative Pre-trained Transformers (GPT), Neural Networks, ARIMA Models, Sentiment Analysis, Cryptocurrency, Hugging Face, Analysis of Variance (ANOVA), APIs, Speech to Text, OCR, Decentralized Finance (DeFi), Trend Analysis, Digital Advertising, Computer Vision, Reinforcement Learning, PEFT, LoRa, Transformers, OpenAI GPT-3 API, OpenAI GPT-4 API, LangChain, Llama 2, Large Language Models (LLMs), Recommendation Systems, FastAPI, Machine Learning Operations (MLOps)
Platforms
Linux, Amazon Web Services (AWS), Docker, Kubernetes, Google Cloud Platform (GCP), Azure
Frameworks
Django, LlamaIndex
Tools
Amazon SageMaker, Bazel
Storage
Redshift, Google Cloud, Elasticsearch
Education
Master's Degree in Data Science
Grenoble Institute of Technology - Grenoble, France
Bachelor's Degree in Computer Science
Grenoble Institute of Technology - Grenoble, France
Certifications
Generative AI with Large Language Models
Coursera
Decentralized Finance (DeFi)
Coursera
AWS Solutions Architect Associate
Pearson VUE
Django for Everybody
University of Michigan | via Coursera
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring