About the Role:
We're looking for a sharp, fast-moving AI/ML engineer who thrives in ambiguity and gets excited about building things from scratch. You'll be tackling greenfield projects across various ML domains - whether that's NLP, time series forecasting, recommendation systems, or computer vision.
The path isn't always clear, and your ability to think on your feet and problem-solve in real-time will be critical. This isn't a role for someone who needs detailed specs and hand-holding. We need someone who can figure it out, move fast, and ship production-quality code.
What You'll Build
- AI/ML systems from the ground up - you'll own projects from conception to production
- Scalable ML pipelines and data workflows
- Production-grade models serving real users at scale
- MLOps infrastructure for training, deployment, and monitoring
- Internal tooling that makes the team more efficient
- Work primarily in the terminal - if you're comfortable in vim/neovim and live in the CLI, you'll fit right in
Required Skills
Core Technical (Non-Negotiable)
- Python - 3-5+ years production experience, this is your primary language
- AI/ML Production - Built and deployed 2-3+ ML models serving real users, not just experiments
- Cloud Platforms - Experience with AWS, Azure, GCP, or OCI for deploying and managing ML workloads. We leverage AI/ML tools across all major cloud providers (Azure AI, AWS SageMaker/Bedrock, GCP Vertex AI, OCI AI Services)
- DevOps - Docker and Kubernetes experience
- Databases - SQL (PostgreSQL, MySQL) and NoSQL/vector databases
- Scripting - Proficient in both Bash and PowerShell for automation
ML Domains (Must have strong experience in at least 2-3 of these)
- NLP/LLMs: Experience with transformers (BERT, GPT, T5), RAG systems, fine-tuning, prompt engineering, or building LLM applications
- Time Series: Forecasting models, anomaly detection, sequential data modeling, or real-time monitoring systems
- Recommender Systems: Collaborative filtering, ranking models, personalization engines, or content recommendations
- MLOps Tools: Production experience with MLflow, Weights & Biases, Kubeflow, Airflow, or similar platforms
- Distributed Training: Large-scale model training, multi-GPU/multi-node setups, efficient data parallelism
Working Style (Critical)
- CLI-first developer - you're comfortable (and prefer) working in the terminal
- Fast thinker - you can rapidly assess problems, prototype solutions, and iterate
- Problem solver - you don't need the answer handed to you; you figure it out
- Greenfield-ready - you're energized by building new things, not just maintaining existing systems
- Self-directed - you can take ambiguous requirements and turn them into working solutions
Nice to Have
- CI/CD Experience: Azure DevOps, GitHub Actions, Jenkins, or similar automation pipelines
- Computer Vision: Production CV experience with PyTorch/TensorFlow, OpenCV, object detection, segmentation, or real-time inference
- Additional Languages: Go or Rust experience for performance-critical components
- Feature stores (Feast, Tecton) or advanced feature engineering
- Model optimization: quantization, pruning, knowledge distillation
- Edge deployment or resource-constrained model deployment
- Experiment frameworks for A/B testing ML models
- Contributions to open-source ML projects
- Real-time streaming data processing (Kafka, Kinesis)
What We're NOT Looking For
- Someone who needs extensive documentation before starting
- Developers who only work with GUIs
- People uncomfortable with ambiguity or rapid change
- Engineers who need constant direction
- Junior developers still learning ML fundamentals
Our Stack
Core: Python | PyTorch/TensorFlow | Scikit-learn | FastAPI/Flask | Git | Bash/PowerShell
ML/AI Tools: MLflow | Airflow/Kubeflow | Azure AI | AWS SageMaker/Bedrock | GCP Vertex AI | OCI AI Services
Infrastructure: Docker | Kubernetes | AWS/Azure/GCP/OCI | PostgreSQL | Azure DevOps | GitHub Actions
Experience Level
- 3-5+ years in AI/ML engineering roles
- Proven track record of shipping 0-to-1 ML projects
- Production ML experience (not just research or coursework)
