Why this role
Engagesoft helps organizations listen continuously, understand what employees are saying in open-ended comments, and turn that into actionable insights. You’ll build the NLP backbone that powers this—at scale and across languages.
What you’ll do
- Understand employee comments at scale: Build models for topic/aspect discovery and clustering within the large embedding spaces of employee comments.
- Select, train, and evaluate embedding models (e.g., BERT/mBERT, RoBERTa, Sentence-Transformers) for clustering, retrieval, classification, and semantic search.
- Ship faithful summarization and evidence-backed narratives that managers can trust.
- Communicate results with stakeholders.
Minimum qualifications
- 3+ years (or equivalent research internships) building NLP systems on real-world text.
- Deep applied understanding of embeddings and transformer architectures (BERT family, Sentence-Transformers).
- Strong Python + ML stack (PyTorch/TensorFlow/JAX), data tooling (Pandas/Numpy), and GPU training/inference.
- Hands-on with two or more: text classification, sequence labeling (NER/PII), semantic search/RAG, topic modeling/clustering, summarization.
- Sound MLOps practices; testing & observability mindset.
- Clear communication and product instincts.
Nice to have
- Multilingual NLP and experience normalizing noisy text.
- Prior work in comment mining (surveys, reviews, tickets) or HR/EX analytics.
- Vector search (FAISS/Milvus), ANN indexing, scalable clustering.
- LLM finetuning/eval (LoRA/QLoRA, preference optimization).
- Advanced degree (Master's or Ph.D.) in Computer Science, Data Science, or a related field.
How we work
- Pragmatic research → shipped services, measured by product impact.
- Tight collaboration with product, data, and design to make insights **explainable and actionable**.
- Continuous improvement with customer-visible A/Bs and quality dashboards.
Compensation & benefits
Competitive compensation aligned to experience and location; details shared during later stages.
Apply
Send your CV, GitHub/Google Scholar (if any), and a short note in a cover letter on an NLP system you took from prototype to production—what worked, what didn’t, and what you’d try next at Engagesoft.
