Role & responsibilities :

We are looking for a hands-on Senior AI/ML Engineer who can own the full lifecycle of machine learning solutions from problem definition and data modelling to training, deployment, monitoring, and continuous improvement. You should be comfortable working with messy real-world data, designing robust data models & features, building and training models, and shipping them to production with proper MLOps practices. You must also be aware of the current AI/ML landscape (LLMs, embeddings, vector search, modern tooling) and know when to use what.

Preferred candidate profile :

Core Technical Skills (6+ Years) :

- Python Programming : Strong expertise in ML libraries (pandas, numpy, scikit-learn, PyTorch, TensorFlow)

- SQL & Databases : Solid SQL skills and hands-on experience with relational and NoSQL data stores

- Production ML : Demonstrated experience shipping end-to-end ML projects to production (not just notebooks / POCs)

- ML Fundamentals : Deep understanding of supervised/unsupervised learning, evaluation metrics, overfitting, bias/variance, data leakage, etc.

MLOps & DevOps :

- Experiment tracking tools (MLflow, Weights & Biases, etc.)

- Model versioning and packaging (Docker, virtualenv, Conda)

- CI/CD pipelines for ML services

- Infrastructure as Code and containerization best practices

Cloud & Architecture :

- Proficiency with at least one major cloud platform :

1. AWS : S3, EC2, SageMaker, Lambda, RDS, DynamoDB

2. GCP : Cloud Storage, Compute Engine, Vertex AI, Firestore

3. Azure : Blob Storage, VMs, Azure ML, Cosmos DB

- API design (REST/GraphQL) and microservice architecture integration

- Understanding of scalability, latency, and cost optimization

Modern AI/ML Landscape Awareness :

- Exposure to LLMs & embeddings (OpenAI, HuggingFace, Anthropic, etc.)

- Familiarity with vector search & semantic search platforms (OpenSearch, Elasticsearch, Pinecone, Weaviate, pgvector)

- Ability to make technical trade-offs between classical ML vs deep learning vs LLM-based approaches

- Understanding of cost, latency, and accuracy considerations for each approach