Akash is currently unavailable
Akash Singh
Verified Expert in Engineering
ML/AI Expert and Developer
Leuven, Belgium
Toptal member since November 25, 2025
Bio
Akash is a senior ML/AI expert with 5+ years of industry and 6+ years of academic experience, holding a doctorate in Trustworthy AI. He specializes in deep learning, deep reinforcement learning (RL), computer vision, NLP, and data science tasks, including churn prediction, fraud detection, classification, and regression. Akash has led teams and delivered full end-to-end model development, deployment, and productionization.
Portfolio
Mohammad Al Mesaeed
Machine Learning, Data Science, Time Series, Time Series Analysis, Python...
University of Liege
AI Research, Deep Reinforcement Learning, Deep Learning, PyTorch, Pandas...
University of Antwerpen
Deep Learning, Deep Reinforcement Learning, Transformers, PyTorch, TensorFlow...
Experience
- Python - 14 years
- Deep Learning - 12 years
- Scikit-learn - 10 years
- PyTorch - 10 years
- TensorFlow - 8 years
- Computer Vision - 6 years
- Deep Reinforcement Learning - 6 years
- Large Language Models (LLMs) - 3 years
Work Experience
Machine Learning Researcher/Data Scientist
2026 - PRESENT
Mohammad Al Mesaeed
- Built and optimized reinforcement learning agents for financial trading, testing multiple RL approaches on different datasets to assess strategy performance, generalization, and stability under varying market conditions.
- Designed and researched reward functions to better align agent behaviour with real market objectives.
- Built scalable research pipelines for offline financial reinforcement learning, supporting structured experimentation and reproducibility comparison of model outcomes across market scenarios.
Technologies: Machine Learning, Data Science, Time Series, Time Series Analysis, Python, Pandas, Linear Regression, Logistic Regression, Decision Trees, Random Forests, K-means Clustering, Support Vector Machines (SVM), Gradient Boosting, Dimensionality Reduction, Reinforcement Learning, Amazon Web Services (AWS), Ubuntu, Docker, Financial Markets, Bitcoin, Algorithmic Trading, Prediction Markets, Cryptocurrency, Jupyter Notebook, Neural Networks, AI Model Training, AI Prompts, GitHub, Multi-armed Bandits (MABs)
AI/ML Engineer | PhD Candidate | Teaching Assistant
2022 - PRESENT
University of Liege
- Developed uncertainty-based ML and RL algorithms for noisy real-world financial institute data.
- Partnered with an insurance company to develop and improve customer churn, acquisition, and retention model by 2.5% under the same budget.
- Built interpretable ML solutions with per-customer risk scores, outperforming existing SOTA baselines.
- Authored four research papers and a technical report. Supervised students and delivered lectures on ML and deep learning.
Technologies: AI Research, Deep Reinforcement Learning, Deep Learning, PyTorch, Pandas, Scikit-learn, Docker, Large Language Models (LLMs), Artificial Intelligence (AI), Machine Learning, Vector Databases, Data Science, Vision Transformer (ViT), Data Analytics, Predictive Analytics, Statistical Analysis, Conversion, Architecture, Reinforcement Learning, Bayesian Inference & Modeling, Open Source, Gradient Boosting, Amazon Web Services (AWS), Agentic AI, Prompt Engineering, Random Forests, Forecasting, XGBoost, Statistical Modeling, Dimensionality Reduction, K-means Clustering, Support Vector Machines (SVM), Logistic Regression, Linear Regression, Classification, Data Analysis, Algorithmic Trading, TradingView, Marketing Analytics, Markov Model, Object Tracking, Computer Vision Algorithms, Prediction Markets, Cryptocurrency, Bitcoin, Jupyter Notebook, Risk Modeling, Risk Models, Image Processing, Neural Networks, Web Scraping, AI Model Training, AI Prompts, Decision Modeling, Back-end, GitHub, Multi-armed Bandits (MABs)
AI Engineer | Researcher
2019 - 2022
University of Antwerpen
- Conducted Imec-funded applied research to enhance state-of-the-art models in computer vision, NLP, deep learning, and RL.
- Developed and evaluated GNNs, relational networks, and transformers for structured data.
- Created and taught RL materials and practicals covering DQN, actor-critic, and PPO.
- Published two research papers on computer vision, RL, and transformers.
Technologies: Deep Learning, Deep Reinforcement Learning, Transformers, PyTorch, TensorFlow, Python, Pandas, Scikit-learn, Computer Vision, Robotics, OpenAI Gym, Graph Neural Networks (GNNs), Databases, AWS IAM, Artificial Intelligence (AI), Machine Learning, Data Science, Vision Transformer (ViT), Source Code Review, Predictive Analytics, Reinforcement Learning, Statistics, Time Series Analysis, Image Recognition, Open Source, Video Analysis, Gradient Boosting, Amazon Web Services (AWS), Object Recognition, OpenCV, TensorFlow Lite, Embedded Linux, Random Forests, Forecasting, Dimensionality Reduction, K-means Clustering, Logistic Regression, Time Series, Classification, Fine-tuning, Data Analysis, Video Processing, Object Detection, YOLOv5, Residual Neural Networks (ResNets), Markov Model, Object Tracking, Computer Vision Algorithms, Jupyter Notebook, Image Processing, Neural Networks, AI Model Training, GitHub, Multi-armed Bandits (MABs)
AI Engineer | Research Scientist
2018 - 2022
Imec
- Developed ML solutions in computer vision and robotics for industrial applications.
- Built and deployed a vision system for Recupel, enhancing recycling efficiency and material recovery across 127,000 tonnes of e-waste with a balanced accuracy of 92% and an AUC of 0.91, maintaining a low Brier score.
- Contributed to applied AI projects under the Flanders Make innovation program.
Technologies: AI Research, MySQL, PostgreSQL, Deep Learning, PyTorch, TensorFlow, Docker, AWS IAM, Azure, Pandas, Deep Reinforcement Learning, Computer Vision, MongoDB, Spark, Artificial Intelligence (AI), Machine Learning, Data Science, Vision Transformer (ViT), Code Review, Source Code Review, Predictive Analytics, Architecture, Time Series Analysis, Image Recognition, Open Source, Gradient Boosting, Amazon Web Services (AWS), Object Recognition, OpenCV, TensorFlow Lite, Embedded Linux, Forecasting, Dimensionality Reduction, Support Vector Machines (SVM), Classification, Fine-tuning, Data Analysis, Video Processing, Object Detection, YOLOv5, Residual Neural Networks (ResNets), Edge AI, Solution Architecture, Robot Operating System (ROS), CI/CD Pipelines, Object Tracking, Computer Vision Algorithms, Jupyter Notebook, Image Processing, Machine Learning Operations (MLOps), Neural Networks, AI Model Training, Decision Modeling, Back-end, GitHub, Multi-armed Bandits (MABs)
Data Scientist
2018 - 2018
Sentiance
- Developed ML models for an IoT behaviour modelling platform.
- Improved venue identification by 27% by integrating GPS with semantic NLP features from open maps.
- Built a classifier for edge case users, e.g., cab drivers, improving classification by 7%.
Technologies: Deep Learning, PyTorch, Pandas, PySpark, Natural Language Processing (NLP), Internet of Things (IoT), Scikit-learn, Python, SpaCy, Artificial Intelligence (AI), Machine Learning, Data Science, Code Review, Source Code Review, Technical Hiring, Interviewing, Task Analysis, Data Analytics, Predictive Analytics, REST APIs, Statistics, Bayesian Inference & Modeling, Gradient Boosting, Random Forests, Forecasting, XGBoost, Dimensionality Reduction, K-means Clustering, Logistic Regression, Classification, Data Analysis, Edge AI, CI/CD Pipelines, Jupyter Notebook, Neural Networks, Web Scraping, AI Model Training, GitHub, Startups
ML Engineer
2017 - 2018
Factmata
- Contributed to the research and implementation of NLP models for misinformation and claim verification.
- Developed and evaluated state-of-the-art classification and semantic similarity pipelines.
- Supported model benchmarking, documentation, and integration into production workflows.
Technologies: Django, MongoDB, PostgreSQL, NGINX, REST, Docker, Python, PyTorch, TensorFlow, Azure, Artificial Intelligence (AI), Machine Learning, Vector Databases, Data Science, Vision Transformer (ViT), Code Review, Source Code Review, Technical Hiring, Interviewing, Task Analysis, SQL, REST APIs, Architecture, Open Source, Amazon Web Services (AWS), Random Forests, Time Series, Classification, Fine-tuning, Solution Architecture, CI/CD Pipelines, Machine Learning Operations (MLOps), Neural Networks, Web Scraping, AI Model Training, Back-end, GitHub, Startups
NLP Scientist
2017 - 2018
Waylo
- Developed conversational AI solutions for the hotel search, pricing intelligence, and conversational booking tools.
- Trained and deployed a hotel-search chatbot with a custom Named Entity Recognition (NER) system.
- Developed and deployed REST APIs powering hotel search and chat management.
Technologies: SpaCy, Python, PyTorch, Pandas, Django, Flask, Artificial Intelligence (AI), Machine Learning, Data Science, Code Review, Source Code Review, Technical Hiring, Interviewing, Task Analysis, Recommendation Systems, SQL, REST APIs, Architecture, Time Series Analysis, Open Source, Amazon Web Services (AWS), Dimensionality Reduction, Time Series, Classification, Fine-tuning, Solution Architecture, Customer Segmentation, Jupyter Notebook, Machine Learning Operations (MLOps), Neural Networks, Web Scraping, AI Model Training, Back-end, GitHub, Startups
Senior Data Scientist
2016 - 2018
ShopClues
- Built large-scale search, recommendation, and personalization systems for eCommerce.
- Optimized models across 27 languages, reducing search latency by 40%.
- Headed a 15-engineer, data science, and applied ML team deploying ranking and personalization to production.
Technologies: Python, Scikit-learn, Elasticsearch, Pandas, SpaCy, Spark, MySQL, Artificial Intelligence (AI), Machine Learning, Data Science, Code Review, Source Code Review, Technical Hiring, Interviewing, Task Analysis, Data Analytics, Predictive Analytics, Statistical Analysis, Recommendation Systems, SQL, Architecture, Gradient Boosting, Object Recognition, OpenCV, Random Forests, Forecasting, XGBoost, Dimensionality Reduction, K-means Clustering, Support Vector Machines (SVM), Logistic Regression, Linear Regression, Classification, Data Analysis, Object Detection, YOLOv5, Residual Neural Networks (ResNets), Solution Architecture, Customer Segmentation, CI/CD Pipelines, Object Tracking, Computer Vision Algorithms, Jupyter Notebook, Image Processing, Neural Networks, AI Model Training, GitHub
Platform Lead | NLP Researcher
2015 - 2016
Neuron
- Developed the NLP platform powering Dainik Bhaskar news recommendation and analytics systems.
- Increased user engagement by 11% and improved ad revenue through relevance optimization.
- Oversaw five engineers developing an analytics dashboard and custom Indian language NER models.
Technologies: Natural Language Toolkit (NLTK), Natural Language Processing (NLP), Computer Vision, NGINX, AWS IoT, Azure, DigitalOcean, Elasticsearch, PostgreSQL, Django, Flask API, Python, SpaCy, Artificial Intelligence (AI), Machine Learning, Vector Databases, Data Science, Code Review, Source Code Review, Technical Hiring, Interviewing, Task Analysis, Statistical Analysis, SQL, REST APIs, Architecture, Statistics, Amazon Web Services (AWS), Object Recognition, OpenCV, Random Forests, Forecasting, XGBoost, Dimensionality Reduction, Support Vector Machines (SVM), Logistic Regression, Classification, Fine-tuning, Solution Architecture, CI/CD Pipelines, Jupyter Notebook, Machine Learning Operations (MLOps), Neural Networks, AI Model Training, TensorFlow, Back-end, GitHub, Startups
Software Developer
2014 - 2015
Algoscale Technologies
- Developed a back-end API and data pipeline for analytics applications.
- Implemented data processing and ML pipelines for healthcare and enterprise reporting.
- Trained and verified different ML models for deployments.
Technologies: Python, Django, Flask, MySQL, Artificial Intelligence (AI), Machine Learning, Predictive Analytics, SQL, REST APIs, Statistics, Random Forests, XGBoost, Linear Regression, Classification, Web Scraping, AI Model Training, Back-end, GitHub, Startups
Project History
Recupel Industrial Computer Vision System | Belgium-wide Impact
https://www.recupel.be/en/blog/recupel-identifies-electronic-devices-using-aiBuilt and deployed a computer-vision system for automated classification of e-waste streams used across Belgium's recycling infrastructure. I achieved a balanced accuracy of 92%, an AUC of 0.91, and strong calibration (low Brier score), improving sorting efficiency and recovery rates across 127,000 tonnes of annual e-waste. The system operationalized industrial robotics and deep learning pipelines with high throughput and reliability.
Uncertainty-aware Churn Prediction for Insurance
Developed uncertainty-aware churn prediction models integrating heterogeneous ensembles and conformal prediction. I improved customer retention by 2.5% without increasing budget, generated per-customer risk scores, and outperformed existing baselines. This work contributes to an uncertainty-aware decision-making pipeline for customer retention and marketing.
Search and Recommendation Engine
Architected and deployed recommendation, ranking, and personalization models for a marketplace serving millions of active users across 27 languages. I reduced search latency by 40%, improved personalization relevance, and managed end-to-end ML lifecycle with a 15-person team.
User Behavioral Modeling and Venue Identification
Developed ML models integrating GPS, accelerometer, and semantic NLP features to classify user movement and venue visits. I improved venue detection accuracy for complex edge-case users, such as cab drivers and mobile workers, by 27%, with a 7% gain in edge-case classification.
Education
2022 - 2025
PhD in Uncertainty in Machine and Reinforcement Learning
University of Liege - Liege, Belgium
2010 - 2014
Bachelor's Degree in Computer Science
Jaypee Institute of Information Technology - Noida, India
Skills
Libraries/APIs
PyTorch, Scikit-learn, Pandas, TensorFlow, XGBoost, REST APIs, OpenCV, PySpark, SpaCy, Natural Language Toolkit (NLTK), Flask API
Tools
GitHub, OpenAI Gym, AWS IAM, AI Prompts, NGINX
Languages
Python, SQL
Platforms
Jupyter Notebook, Docker, Amazon Web Services (AWS), Microsoft, Linux, Azure, AWS IoT, DigitalOcean, Ubuntu, Embedded Linux
Storage
Microsoft SQL Server, Databases, PostgreSQL, Azure SQL, MySQL, MongoDB, Elasticsearch
Frameworks
Django, TensorFlow Lite, Multi-armed Bandits (MABs), Spark, Flask
Paradigms
Object-oriented Programming (OOP), REST
Other
Deep Learning, Large Language Models (LLMs), Explainability, AI Research, Transformers, Computer Vision, Artificial Intelligence (AI), Machine Learning, Data Science, Code Review, Source Code Review, Technical Hiring, Task Analysis, Data Analytics, Predictive Analytics, Image Recognition, Decision Trees, Gradient Boosting, Object Recognition, Random Forests, Forecasting, Support Vector Machines (SVM), Classification, Data Analysis, Customer Segmentation, Object Tracking, Computer Vision Algorithms, Image Processing, Neural Networks, Web Scraping, AI Model Training, Decision Modeling, Back-end, Deep Reinforcement Learning, APIs, Robotics, Natural Language Processing (NLP), Vector Databases, Vision Transformer (ViT), Interviewing, Statistical Analysis, Conversion, Architecture, Reinforcement Learning, Statistics, Time Series Analysis, Open Source, Video Analysis, Agentic AI, Statistical Modeling, Dimensionality Reduction, K-means Clustering, Logistic Regression, Linear Regression, Fine-tuning, Video Processing, Object Detection, YOLOv5, Residual Neural Networks (ResNets), Blob Storage, Marketing Analytics, Risk Management, Risk Modeling, Risk Models, Machine Learning Operations (MLOps), Startups, AI Agents, Graph Neural Networks (GNNs), Internet of Things (IoT), Retrieval-augmented Generation (RAG), Recommendation Systems, Bayesian Inference & Modeling, Financial Markets, Prompt Engineering, Time Series, Edge AI, Solution Architecture, Robot Operating System (ROS), Algorithmic Trading, TradingView, Markov Model, CI/CD Pipelines, Prediction Markets, Cryptocurrency, Bitcoin, Trading
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