Machine Learning Engineer - Computer Vision, LLM & GenAI

Core Responsibilities

Model Development & GenAI
- Build & deploy computer vision models for image classification, object detection, segmentation
- Develop LLM-based solutions for text analysis, content generation, information extraction
- Design & implement agentic AI workflows for autonomous decision-making & multi-step reasoning
- Prompt engineering & optimization for domain-specific LLM tasks
- Implement core ML algorithms with full understanding (not just library usage)
- Fine-tune foundation models for specialized use cases

Algorithm & Model Expertise
- Deep understanding of ML algorithms (supervised, unsupervised, reinforcement learning)
- Model evaluation, validation, and performance optimization
- Hyperparameter tuning and experimentation frameworks
- Bias detection and model fairness assessment

MLOps & Cloud-Native Development
- Deploy models to production across AWS & Google Cloud Platform ecosystems
- Implement model versioning, A/B testing, performance tracking
- Ensure model governance, reproducibility & compliance
- Optimize cloud infrastructure for cost efficiency

Required Skills

Technical (Must-Have)
- Python (pandas, scikit-learn, PyTorch/TensorFlow, OpenCV)
- Computer Vision Image classification, object detection, segmentation, feature extraction
- GenAI & LLM Prompt engineering, RAG (Retrieval-Augmented Generation), fine-tuning, embeddings
- Agentic AI:Multi-step reasoning, tool use, agent frameworks (LangChain, AutoGen, CrewAI)
- ML algorithms from scratch (not just library usage)
- Statistical analysis & experimental design

Cloud-Native & MLOps (Must-Have)
- AWS SageMaker (training, endpoints, pipelines), Bedrock, Lambda, EC2, S3, EFS, Glue, CloudWatch
- Google Cloud Platform Vertex AI (AutoML, custom training, Generative AI APIs), Compute Engine, Cloud Storage, App Engine, Cloud Run
- Containerization & orchestration (Docker, Kubernetes basics)
- Infrastructure-as-Code (Terraform, CloudFormation)
- Cost monitoring & optimization across cloud platforms
- Logging, monitoring, alerting (CloudWatch, Cloud Logging, Prometheus)

Nice-to-Have
- Domain-specific expertise (Auto, healthcare, finance, retail, etc.)
- Explainable AI (SHAP, LIME, attention visualization)
- Web dashboards & visualization (Streamlit, Dash, Plotly)
- SQL for data pipelines & ETL
- CI/CD pipelines (GitHub Actions, Cloud Build)
- Vector databases (Opensearch, Pinecone, Weaviate, Milvus) for RAG
- LLM evaluation frameworks (RAGAS, DeepEval)
- Model monitoring & drift detection

Key Deliverables

Scalable ML models with documented accuracy metrics (precision, recall, F1, AUC, etc.)  

Cost-optimized cloud infrastructure (spot instances, auto-scaling, resource right-sizing)  

GenAI-powered workflows (LLM chains, agents, multi-step reasoning)  

Interactive dashboards & demos for stakeholders  

Production-ready, maintainable code with comprehensive monitoring  

Model interpretability & explainability documentation  

Automated ML pipelines (SageMaker Pipelines / Vertex AI Pipelines)  

Agentic AI systems for autonomous decision-making  

Reproducible experiments & model versioning strategy