AI Engineer

As an AI Engineer at LENSEC, you will design, build, and deploy AI solutions as part of a production-grade AI platform. This role is hands-on and focused on developing, training, and integrating models into real-world systems, working closely with data, backend, and product teams.

This role focuses on applied machine learning for computer vision use cases, as well as generative AI development — LLM-powered features, agentic workflows, and the retrieval and orchestration layers around them — with a strong emphasis on inference efficiency and cost optimization in production systems.

Job Responsibilities

• Design, develop, and optimize computer vision models for tasks such as detection, classification, tracking, and behavior analysis.
• Prepare and manage datasets for training and evaluation, including data cleaning, labeling strategies, and augmentation.
• Train, fine-tune, and evaluate machine learning and deep learning models with a focus on accuracy, robustness, and performance.
• Integrate computer vision models into backend services and pipelines for inference and downstream processing.
• Optimize models for inference efficiency (latency, throughput, resource usage) in production environments.
• Implement model evaluation, validation, and monitoring to ensure consistent performance over time.
• Collaborate with backend and frontend engineers to expose model outputs via APIs/docker and user-facing features.
• Document model assumptions, limitations, and performance characteristics.
• Design and build generative AI features using large language models — prompting, structured output, function/tool calling, and evaluation of model behavior.
• Develop agentic workflows that combine LLMs with tools, retrieval, and external systems to complete multi-step tasks reliably.
• Build and maintain retrieval-augmented generation (RAG) pipelines, including chunking, embeddings, vector search, and context assembly.
• Fine-tune or adapt foundation models where appropriate (e.g., instruction tuning, LoRA/QLoRA) and measure the gain against prompting baselines.
• Optimize LLM inference for latency, throughput, and cost — model selection and routing, caching, batching, quantization, and token budgeting.
• Define and run evaluation for generative systems, including offline benchmarks, regression suites, and guardrails for accuracy, safety, and hallucination.
• Stay current with advances in computer vision and applied machine learning.

Qualifications

• 3+ years of professional experience in machine learning, computer vision, or applied generative AI roles.
• Demonstrated experience delivering computer vision models into production or pilot environments.
• Experience shipping at least one generative AI or LLM-based feature into production or a pilot environment.
• Strong problem-solving skills and ability to translate real-world problems into ML solutions.

Technical Requirements

• Programming Languages: Python (primary), with strong software engineering practices
• Computer Vision: Solid understanding of classical and deep learning–based computer vision techniques
• ML Frameworks: Experience with at least one major deep learning framework (e.g., PyTorch or TensorFlow)
• Data Handling: Experience working with image/video datasets and data preprocessing pipelines
• Model Training: Experience training and fine-tuning deep learning models
• Evaluation: Knowledge of model evaluation metrics and validation techniques
• Integration: Experience integrating ML models into production systems or services
• Generative AI: Hands-on experience building applications on large language models (API-based or self-hosted)
• Prompt & Context Engineering: Practical skill with prompting, structured outputs, and tool/function calling
• Agents: Experience with agentic patterns — tool use, planning, multi-step orchestration, and failure handling
• Retrieval: Working knowledge of embeddings, vector databases, and RAG pipeline design
• LLM Evaluation: Ability to define task-specific metrics and build evaluation sets for generative outputs
• Inference Optimization: Experience reducing latency and cost for model serving (caching, batching, quantization, model routing)

Qualified Candidates can send their CVs to : [Confidential Information]