AI\NLP Engineer

Are you an Algorithms Developer looking to influence the world of English Language

Evaluation? In a world of assessments which can be outsmarted by guesswork and

prefab answers, www.speaknow.co is the only solution in the language testing space

to measure real life communication capabilities by evaluating natural speech and

writing.

Using a combination of cutting-edge technology and advanced linguistics research,

the Speaknow test is designed to measure English quickly and accurately. Our test is

used by leading organizations around the world to assess English abilities in

recruitment and language teaching, and is trusted by educational institutions and

government offices to qualitatively measure English abilities in a wide variety of

cases.

What will you do with us?

Speaknow is looking for a hands-on, execution-oriented AI/NLP Software Engineer

to join our dynamic R&D team and take ownership of our ML models used across our

flagship assessment products.

Your work will involve close collaboration with linguists and psychometric experts,

and requires strong engineering discipline, independent problem-solving, and full

end-to-end ownership.

The position will involve building and improving real-world NLP and speech-scoring

systems that run at scale and impact global customers.

Experience, qualifications & skills we are looking for:

• Design, develop and improve NLP and Machine Learning based algorithms.

• Develop our main platforms using the latest technologies and machine learning.

• Handling development cycles, including architecture, training pipelines,

deployment, and ongoing monitoring.

• Provide innovative solutions for large-scale data and traffic.

• Own ML models end-to-end, from data preparation to production deployment.

• Work with ASR, embeddings, prosodic features, and transformer-based models to

improve scoring accuracy and reliability.

Requirements:

• BS, MS or Ph.D. degree in Computer Science, or equivalent field.

• 3+ years Industry experience specializing in machine learning techniques and

concepts including Deep Learning, regression, and classification models.

• Proven experience fine-tuning and optimizing transformer architectures in

production (BERT, T5, Llama, Mistral, GPT-style models).

• Experience with text analysis, sequence modelling, and semantic feature extraction.

• Hands-on experience building training pipelines, deploying models to production,

and monitoring performance.

• Experience with speech analysis pipelines (ASR, segmentation, feature extraction,

prosodic analysis).

• High standards for code quality, testing and performance.

• High-level interpersonal skills.

• Proficient English.

• Strong ability to work independently and solve ambiguous, open-ended problems.

Advantages (Nice to have):

• Experience building scoring or assessment models in ed-tech or evaluation contexts.

• Background in computational linguistics, psychometrics or language-testing

research.

• Experience optimizing models for latency, cost and scale in production

environments.

• Familiarity with ASR, speech scoring, prosodic analysis or signal processing.

• Knowledge of statistical reliability (ICC, Cronbach’s alpha) and their use in

assessment validation.

• Experience working with large-scale inference systems and real-time scoring

workloads.