Focus: Computer Vision, OCR, NLP, Search & Production ML Systems
Contract: B2B
English: C1+
Compensation: Gross, negotiable
Public Holidays: 10 per year (vacation & sick leave unpaid)

1. Mission of the Role

Build and operate production-grade AI systems that convert complex, unstructured technical content into high-quality structured data and power fast, accurate search.

This is not a research-only position.
You will own the full lifecycle: architecture → experimentation → deployment → observability → continuous improvement.

2. What You Will Own

A. End-to-End ML Architecture

• Design hybrid systems combining:

  • Computer Vision (detection, segmentation)

  • Layout understanding

  • OCR pipelines

  • NLP processing

• Transform unstructured documents into structured, searchable representations.

• Design hybrid retrieval strategies (dense + sparse ranking).

B. Production ML Systems

• Ship containerized ML services to AWS.

• Build scalable inference microservices.

• Optimize pipelines for:

  • Latency

  • Throughput

  • Cost

  • Reliability

• Define and maintain service-level objectives (SLOs).

C. Search & Relevance Engineering

• Implement embedding pipelines.

• Design hybrid search ranking logic.

• Work with vector databases and similarity search.

• Improve recall and precision using measurable experiments.

D. MLOps & Operational Excellence

• Implement reproducible ML workflows:

  • Model & dataset versioning

  • CI/CD for ML

  • Automated validation

  • Rollout / rollback strategies

• Build observability using:

  • OpenTelemetry

  • Prometheus

  • Grafana

• Set up dashboards and alerting for:

  • Model drift

  • System degradation

  • Latency spikes

E. Evaluation & Quality Control

• Design “Gold Standard” datasets and labeling guidelines.

• Define measurable quality metrics and acceptance thresholds.

• Conduct systematic error analysis.

• Maintain prioritized improvement backlogs based on impact.

3. Required Background

Experience

• 7+ years backend engineering (Python).

• 5+ years hands-on ML engineering in production.

• Proven track record of deploying ML systems into real-world environments.

Computer Vision

• Object detection

• Image segmentation

• OCR pipeline design

• Layout-aware document processing

NLP

• Information extraction

• Intent recognition

• Annotation parsing

• Structured data transformation

ML & Engineering Stack

• PyTorch (preferred), TensorFlow, scikit-learn

• Strong software engineering fundamentals:

  • Clean architecture

  • Testing

  • Code review

  • Maintainable design

Search & Retrieval

• Vector databases

• OpenSearch / Elasticsearch

• Dense + sparse hybrid search strategies

Cloud & Infrastructure

• AWS

• Docker

• Kubernetes

• CI/CD pipelines

• Microservices architecture

Production Operations

• Observability and alerting systems

• High-compliance / secure environments

• Data isolation requirements