15 Nov
|
Coxvi Technologies
|
Bengaluru
15 Nov
Coxvi Technologies
Bengaluru
JOB DESCRIPTION
AI Engineer Computer Vision & LLM Systems Were looking for an AI Engineer with strong fundamentals, sound engineering judgment, and the curiosity to solve difficult, open-ended problems. Youll build AI systems that understand documents, images, and technical information. The work spans computer vision, document intelligence, retrieval-augmented generation (RAG), and LLM-powered workflows
- from early experiments to reliable production features. What you’ll work on
- Build systems to extract and interpret information from complex PDFs, scanned documents, tables, diagrams, and engineering drawings.
- Solve visual understanding problems involving layout, symbols, objects, and relationships using computer vision and multimodal models.
- Develop and improve RAG pipelines, including ingestion, chunking, embeddings, retrieval, reranking, and grounded answer generation.
- Build LLM workflows with structured outputs, tool calling, and agents where they add practical value.
- Investigate failures, identify root causes, and improve accuracy through systematic experimentation.
- Create evaluation datasets and benchmarks to measure quality, latency, and cost.
- Integrate models into backend services and own their behavior in production. What we value We care about depth of understanding and demonstrated problem-solving ability. You should be able to explain why you chose an approach, where it fails,
how you evaluated it, and what you would try next. What we’re looking for
- Strong foundations in machine learning and deep learning, including training, inference, overfitting, generalization, and evaluation.
- A working understanding of the mathematics behind common ML methods: linear algebra, probability, statistics, and optimization.
- Solid Python skills and the ability to write clear, maintainable, testable code.
- Practical experience with computer vision or document understanding
- for example, OCR, object detection, layout analysis, or vision-language models.
- Understanding of transformers, embeddings, LLM capabilities and limitations, and retrieval systems.
- The ability to break an unfamiliar problem into testable hypotheses, establish a baseline, and improve it using evidence.
- Ownership: you can take a problem from investigation through implementation and validation, and communicate tradeoffs clearly. Good to have
- Experience with PyTorch, OpenCV, Hugging Face, or similar tools.
- Experience extracting information from technical drawings, schematics, or P&IDs.;
- Experience with vector databases, hybrid search, and reranking.
- Experience fine-tuning models, curating datasets, or working with limited labeled data.
- Experience deploying AI services using APIs, Docker, and cloud infrastructure.
📌 AI Engineer (Bengaluru)
🏢 Coxvi Technologies
📍 Bengaluru