Lead Data Scientist at Bengaluru (Bangalore) Full Time

The DataFlow Group is a leading global provider of specialized Primary Source Verification (PSV) solutions, background screening and immigration compliance services. The DataFlow Group partners with clients across the public and private sectors to assist them in mitigating potential risk by exposing fraudulent:

Duties and responsibilities:

Key Responsibilities:

● Model Selection & Training:

○ Lead the process of selecting and training state-of-the-art machine

learning models tailored to specific computer vision tasks, including object detection, image segmentation, and anomaly detection.

○ Apply advanced techniques for model fine-tuning and optimization to

achieve the best performance for various applications.

○ Leverage your expertise in YOLO, ViT (Vision Transformers), and other popular architectures to develop high-performing models.

● Computer Vision & Image Processing:

○ Develop and implement computer vision algorithms for tasks such as

object detection, image segmentation and anomaly detection.

○ Use tools like OpenCV for image pre-processing, augmentation, and other necessary transformations for robust model training.

● Model Deployment & Serving:

○ Deploy machine learning models in production environments using

TensorFlow or PyTorch Serving to create highly scalable and performant inference pipelines.

○ Ensure smooth integration of models with production systems and

optimise them for latency, throughput, and memory efficiency.

● MLOps & Data Pipelines:

○ Build and maintain end-to-end machine learning pipelines, from data

ingestion and processing to model inference and monitoring.

○ Apply best practices in ML Ops for versioning, tracking experiments,

automating workflows, and managing deployments.

○ Work with cloud platforms (AWS, GCP, Azure) and containerization

technologies (Docker, Kubernetes) for model serving.

● SQL & Data Management:

○ Write optimised and efficient SQL queries to extract and manipulate large datasets for model training and evaluation.

○ Analyse structured and unstructured data to generate insights and

features for model development.

● Model Evaluation & Performance Metrics:

○ Use statistical methods and metrics to evaluate model performance,

ensuring high accuracy, precision, recall, F1 score, etc., on real-world tasks.

Skills/ Qualifications:

1. Qualifications:

a. Bachelors or Masters. in Computer Science, Machine Learning, Artificial Intelligence, or related field.

b. 5+ years of hands-on experience in data science and machine learning, with a focus on computer vision.

2. Technical know-how:

a. Strong expertise in computer vision tasks, including object detection (YOLO), image segmentation, and image classification.

b. Knowledge of using tools like LabelMe to create image annotations required for training data labelling.

c. In-depth knowledge of deep learning frameworks such as TensorFlow, PyTorch, and Keras.

d. Familiarity with Vision Transformer (ViT) models and their applications.

e. Proficiency with OpenCV for image pre-processing, augmentation, and other related tasks.

f. Experience in building and deploying machine learning models using

TensorFlow Serving, PyTorch Serving, or similar tools.

g. Strong MLOps experience, with hands-on knowledge of creating and

maintaining data pipelines, automating workflows, and managing model

deployments using tools like Docker, Kubernetes, and cloud platforms (AWS, GCP, Azure).

3. Data Handling:

a. Expertise in writing optimised SQL queries for large-scale data processing and analysis.

b. Ability to work with large, complex datasets and efficiently query, clean, and process them for training and inference purposes.

4. Model Evaluation:

a. Strong understanding of statistical methods and metrics to evaluate machine learning model performance.

b. Experience with A/B testing, model validation techniques, and performance benchmarking.

Good to have:

● Experience with large-scale distributed machine learning systems.

● Knowledge of advanced techniques such as self-supervised learning, transfer learning, or reinforcement learning.

● Familiarity with data versioning and experiment tracking tools like MLflow or DVC.

● Experience in integrating machine learning models into production environments for real-time inference.