The Embedding API analyzes an HR object (a profile, a job, an experience, an education, a list of skills, a summary, a project, and more) and returns a numerical vector representing the input HR object in a 1024-dimensional space.
![embedding image](/static/40258de5904324a8ce5e2ab9632c4b1d/4cf73/embedding.png)
Embedding API
Build faster great AI algorithms for HR.
Accelerate 10x your R&D roadmap. Overcome HR data biases. Achieve high-performance with even a tiny amount of HR data.
Focus on delivering high-quality
AI algorithms and not features engineering
The vectors of similar HR objects will be close to each other in the 1024-dimensional space. Therefore, AI developers can use the Embedding API for filtering, indexing, ranking, and organizing HR objects according to semantic similarity.
![Broad use cases](/static/2938548496ddf3fbe4bd5753cbb2bbda/2b3ff/book.png)
Broad use cases
Similarity analysis, Search and retrieval, Machine transfer learning.
![Multiscale modeling](/static/ed1aa14fa5974990d7fb4d7e423f51e7/2b3ff/sitemaps.png)
Multiscale modeling
A large portfolio of vector families to satisfy your use cases: Profile2Vec, Job2Vec, Experience2Vec, Education2Vec, Skills2Vec, Interests2Vec.
![Representation debiasing](/static/55887e56a72069334bed143175a20fa0/2b3ff/balance.png)
Representation debiasing
Fairness is not the default; that‘s why we built inclusive models that measure and mitigate unintended bias in the HR data.
![Words N-gram](/static/4c65ffe1893e9770307690fe0fad7893/2b3ff/puzzle.png)
Words N-gram
Supports sequences of adjacent words that have a sense together.
![Sub-word information](/static/4c5dcdaf7984b6dcdc6a8a5d61a6c1af/2b3ff/dice.png)
Sub-word information
We enrich word vectors with relevant sub-word information.
![Multiple Languages](/static/2bb63b3db38322f4454f414cd23cdce4/2b3ff/language.png)
Multiple Languages
Supports 32+ languages (Arabic, Chinese, Dutch, English, French, German, Italian, Japanese, Portuguese, Romanian, Spanish, and more.)
Your return on investment
![Average accuracy](/static/d689108c970ba5f3a7cfdd26bcc26817/1db4f/arrow-chart.png)
95%
Average accuracy
To solve your HR challenges
![Languages](/static/2bb63b3db38322f4454f414cd23cdce4/567d1/language.png)
32+
Languages
To scale globally
![R&D acceleration](/static/d4bb8ef61d3d0c22ae37a489bc75b9be/1db4f/clock.png)
10x
R&D acceleration
To focus on impact
The HrFlow.ai Embedding API lifts a heavy workload on your R&D team
Benefit from 6 years of expertise in representing HR objects.
Embedding
TF-IDF
L.D.A
Doc2Vec
Seq2Seq
Transformer
HrFlow.ai
N-grams
Multilingual
Sub-words
Dimensionality
reduction
reduction
Compression
Neural
memory
memory
Semantic
Quantization
Unicode
Allocation
bias
bias
Representation
bias
bias
Multi-scale
modeling
modeling
Hierarchical
representation
representation
Trained on
HR Data
HR Data
Don't take our word for it
96%
Job matching accuracy
![Derek Chang photo](/static/5f3861773f0f569d12bea8017c7889de/f9fdc/photo.png)
« HrFlow.ai Embedding API allowed us to build quickly a bias-free staffing AI algorithm that has surpassed our historical human performance. »