About CONXAI
CONXAI has built a no-code, agentic AI platform for the Architecture, Engineering and Construction (AEC) and physical industries, focused on knowledge-automation. We automate high-stakes, knowledge-intensive workflows traditionally trapped in siloed data, fragmented tools and tacit (undocumented) human expertise.
Our multi-agent systems perform complex reasoning in the physical world; and transform bespoke, service-heavy processes into scalable Service-as-a-Software automation.
CONXAI is trusted by some of the leading AEC companies in Europe, US, LATAM and Japan.
Your Role
You bridge the gap between SOTA research and real-world deployments. You are responsible for ensuring Computer Vision models perform reliably when exposed to complex, unstructured customer data.
Core Responsibilities
- Own the Feedback Loop: Monitor production data to identify exactly where and why models struggle in specific customer environments
- Diagnose & Propose: Analyze discrepancies between model output and reality to propose concrete algorithmic or data-driven fixes
- Continuous Validation: Own the "last-mile delivery" by ensuring proper use-case setup and validating the accuracy of final results for the customer
- Drive Data-Centric Improvements: Lead the data "flywheel" by curating specialized datasets and integrating high-value customer data for model retraining
- Operationalize SOTA: Adapt high-level architectures into performant, cost-effective solutions tailored for specific customer use-cases
- Validate for Impact: Design evaluation frameworks that measure true customer value rather than relying solely on standard benchmarks
What We’re Looking For
- Bachelor's / Master's degree in Computer Science (or related) or Civil Engineering with specialization in Data Science / Lean Construction
- Experience with training, and evaluating Deep Learning models in PyTorch
- Good understanding of basics in machine learning and computer vision, specifically representation learning
- Experience training and fine-tuning a Deep Learning architecture for:
- Object Detection
- Segmentation
- Large Vision Language Models
- Experience with data curation, visualization, outlier detection, active learning and hard-negative mining
- Proficient in Python and good software engineering skills
- Ability to work in a team-oriented environment
- Strong problem-solving skills
- Good communication and interpersonal skills (fluent in English)
Why CONXAI
- Edge of Innovation: Be at the absolute forefront of AI in the construction tech space
- High Autonomy: Contribute to a new paradigm for multi-modal scene understanding and reasoning - owning the logic, performance, and customer impact
- Top-Tier Peer Group: Work with a global team of ML engineers, software engineers and industry practitioners
- Equity & Scale: Competitive compensation with significant equity upside
Skills Required
- Bachelor's or Master's degree in Computer Science or related field, or Civil Engineering with Data Science/Lean Construction specialization
- Experience training and evaluating deep learning models in PyTorch
- Strong understanding of machine learning and computer vision fundamentals, specifically representation learning
- Experience training and fine-tuning architectures for object detection
- Experience training and fine-tuning architectures for segmentation
- Experience with large vision-language models
- Experience with data curation, visualization, outlier detection, active learning and hard-negative mining
- Proficiency in Python and strong software engineering skills
- Ability to work in a team-oriented environment
- Strong problem-solving skills
- Good communication and interpersonal skills; fluent in English
What We Do
CONXAI provides a no-code, agentic AI platform purpose-built for the architecture, engineering, and construction (AEC) industry. The platform automates knowledge-intensive workflows by transforming fragmented project data and tacit human expertise into structured, actionable insights. By integrating diverse data sources, CONXAI enhances project control, improves operational efficiency, and prevents knowledge loss within the construction sector.
