LimeoX is looking for a Senior NLP & Machine Learning Systems Engineer to join the Platinum tier of our engineering organization.
This is not a conventional AI development role.
You will work at the intersection of NLP, machine learning, real-time inference, financial data, and large-scale ML systems — solving problems where prediction quality, data quality, and processing speed all matter.
02 — What you'll work on
- Real-time financial news and event intelligence
- Market-moving event detection and classification
- Financial NLP and information extraction
- Direction, magnitude, and confidence prediction
- Macro-economic event analysis
- Large-scale training and labeling datasets
- Model evaluation, experimentation, and continuous improvement
The objective isn’t simply to build models. It’s to build ML systems capable of turning information into intelligence fast enough to matter.
03 — What we're looking for
We want engineers who have already gone significantly beyond tutorials, notebooks, and API integrations.
Machine learning
- Designing, training, evaluating, and improving production ML models
- PyTorch, TensorFlow, or equivalent frameworks
- Classification, ranking, embeddings, or predictive modeling
- Handling large, noisy, and highly imbalanced datasets
- Precision/recall/F1, calibration, threshold optimization, and robust evaluation
- Dataset design, leakage prevention, and experimentation
NLP
- Transformers and modern NLP architectures
- Text classification and information extraction
- Named entity recognition / entity linking
- Semantic similarity and embeddings
- Designing NLP pipelines beyond simple LLM prompting
ML systems
- Production model deployment and inference
- GPU-based workloads and optimization
- Batch and real-time processing
- Building scalable ML/data pipelines
- Profiling and reducing inference bottlenecks
- Monitoring production models and data quality
04 — You should be able to answer questions like
- Q.01 How would you identify a genuinely abnormal event among hundreds of thousands of financial news items?
- Q.02 How would you prevent future information from leaking into a financial ML training dataset?
- Q.03 How would you determine whether a model with 95% accuracy is actually useless?
- Q.04 How would you reduce an NLP pipeline from 20 seconds of processing time to 2 seconds?
- Q.05 How would you evaluate whether an event-prediction model works outside the market regime on which it was trained?
We care considerably more about your answers to questions like these than how many AI tools you can list on your CV.
05 — What this role is not
This role is not intended for entry-level or early-career AI candidates.
If most of your AI experience consists of:
- Building RAG/chatbot applications
- LangChain/LlamaIndex integrations
- University ML assignments
- Following online ML tutorials without substantial experience building, training, evaluating, and deploying your own ML systems
then this role is unlikely to be the right fit.
06 — What strong candidates may have done
There is no single required background.
You might have:
- Built production NLP/ML systems processing millions of records
- Trained or fine-tuned transformer-based models
- Designed large-scale ML datasets and evaluation pipelines
- Optimized GPU inference or production ML latency
- Conducted applied ML research
- Built financial, recommendation, search, fraud, forecasting, ranking, or event-detection systems
- Published meaningful ML/NLP research
- Taken ML systems from experimentation into production
Education A Master’s or PhD in Computer Science, AI, Machine Learning, Data Science, Statistics, Mathematics, or a related field is highly valued — but exceptional engineering evidence can outweigh credentials.