Are you a Computer Vision hero? Then we are looking for you!

Design and implement robust computer vision pipelines for real-world image and video processing tasks. Work across edge and cloud environments to deliver performant, scalable, and production-ready vision systems under real-world constraints.

KEY RESPONSIBILITIES

General responsibilities include:

• Develop and optimise image and video processing pipelines using OpenCV, PyTorch, and TensorFlow.
• Implement deep learning models for object detection, tracking, segmentation, OCR, and pose estimation.
• Build scalable inference systems for edge (Jetson, Coral, mobile) and cloud (AWS, Azure) deployments.
• Design and maintain data pipelines for video ingestion, annotation, augmentation, and model training.
• Integrate vision models into production systems with real-time or near-real-time performance.
• Collaborate with hardware and backend teams to optimise latency, throughput, and memory usage.
• Conduct error analysis and model evaluation using custom metrics and visualisation tools.
• Ensure robustness across lighting, occlusion, motion blur, and domain shifts.
• Contribute to CI/CD pipelines, model versioning, and reproducibility practices.
• Document model architectures, training procedures, and deployment workflows.

EXPERIENCE + QUALIFICATION

• 3-6 years of experience in computer vision and deep learning.
• Strong proficiency in Python and libraries like OpenCV, PyTorch, TensorFlow, and scikit image.
• Experience with real-time video processing, edge deployment, and hardware acceleration (CUDA, TensorRT).
• Solid understanding of CNNs, transformers, and vision-specific architectures (YOLO, Mask R-CNN, SAM).
• Familiarity with MLOps tools (MLflow, Weights & Biases) and model lifecycle management.
• Strong debugging, profiling, and optimisation skills for vision workloads.
• Commitment to code quality, reproducibility, and performance benchmarking

NICE TO HAVES

• Experience with OCR pipelines, document layout analysis, or keypoint-based pose estimation.
• Experience with AWS or Azure for model training and deployment.
• Familiarity with synthetic data generation, augmentation strategies, and domain adaptation techniques.
• Exposure to multi-modal models (e.g., CLIP, DINO) and vision-language alignment.
• Knowledge of camera calibration, lens distortion correction, and photometric consistency.

WORK DETAILS

This is a full-time position, where you’ll be working in a supportive, fast-paced environment with access to learning and development opportunities.

• Type: Full-time, onsite with some remote flexibility
• Schedule: Standard 8 hours, five-days engagement under contract
• Location: Acme AI Ltd, Level 5, House 385, Road 6, Mirpur DOHS, Dhaka 1216
• Salary: BDT 60,000-110,000 (offered based on experience and technical test-result)

BENEFITS

• Competitive salary based on experience
• Festival bonuses
• Access to company health fund (upon eligibility; since health and hospitalisation insurance sucks!)
• Quarterly profit-sharing (upon eligibility)

How to apply?

Please lodge your applications by filling out  this Google form . Application Deadline:  Considered on a rolling basis until 4 December 2025 at 11.59 PM