GCP, BigQuery, Python

Job Summary

Results-driven ML/AI Engineer with expertise in designing and deploying end-to-end machine learning pipelines from data ingestion and feature engineering to model deployment and monitoring. Skilled in building scalable ML solutions optimizing models and implementing robust MLOps practices using modern tools and cloud platforms. Experienced in developing advanced Generative AI and LLM-based systems including RAG architectures prompt engineering and AI agents.

Responsibilities

Proficient in ensuring model accuracy safety and performance through rigorous evaluation testing and monitoring frameworks. Strong background in data engineering real-time integration and enterprise system connectivity with a focus on delivering reliable high-quality AI solutions. Collaborative team player with experience in Agile environments stakeholder communication and mentoring junior engineers.

Core ML Engineering

Design and implement end-to-end ML pipelines data ingestion feature engineering model training evaluation and deployment

Train fine-tune and optimize models (XGBoost LightGBM PyTorch TensorFlow) for classification regression clustering and ranking tasks

Build and maintain feature engineering pipelines using Spark Pandas and Feast

Deploy models as REST APIs (FastAPI Flask) and batch inference jobs

Implement model versioning experiment tracking and reproducibility using MLflow and Weights & Biases

Monitor model performance in production drift detection retraining triggers rollback procedures

Write unit tests integration tests and regression test suites for ML components

Generative AI & LLM Engineering

Design and implement RAG (Retrieval-Augmented Generation) architectures chunking strategies embedding pipelines hybrid retrieval (BM25 + vector) reranking

Build and optimize LLM prompt pipelines using chain-of-thought few-shot and system prompt techniques

Integrate foundation models via APIs OpenAI (GPT-4o) Anthropic (Claude) IBM Granite Meta LLaMA Mistral

Fine-tune large language models using SFT LoRA QLoRA RLHF and DPO techniques

Build multi-LLM orchestration systems routing tasks by cost latency and capability

Implement AI agents with tool use memory management and multi-step reasoning (LangChain LlamaIndex AutoGen)

Design and maintain vector database pipelines Pinecone Weaviate pgvector ChromaDB

Deploy LLM inference services with low-latency optimization using vLLM TGI Triton Inference Server

Accuracy Safety & Evaluation

Implement hallucination detection faithfulness checks and PII redaction guardrails before LLM output delivery

Evaluate LLM output quality using RAGAS FactScore BERTScore and LLM-as-judge frameworks

Build evaluation pipelines tracking precision recall F1 PR-AUC and domain-specific metrics

Conduct A/B testing and shadow deployment for model version comparison

Perform bias and fairness audits across demographic subgroups

Build automated regression test suites to catch model degradation before production release

MLOps & Infrastructure

Build CI/CD pipelines for model training testing and deployment using GitHub Actions Jenkins and Azure DevOps

Orchestrate ML workflows using Apache Airflow Kubeflow Pipelines and Prefect

Containerize ML workloads with Docker and deploy on Kubernetes with GPU autoscaling (HPA KEDA)

Set up model serving infrastructure using KServe BentoML and Ray Serve

Implement feature stores (Feast Tecton) for consistent online and offline feature serving

Monitor inference latency throughput and error rates using Prometheus Grafana and Datadog

Manage cloud ML platforms AWS SageMaker Azure ML GCP Vertex AI IBM watsonx.ai

Data Engineering & Integration

Build data ingestion pipelines from structured (SQL APIs) and unstructured (PDFs emails documents) sources

Process and transform large-scale datasets using Apache Spark Databricks and dbt

Design document parsing pipelines supporting PDF DOCX HTML and JSON for RAG ingestion

Integrate AI/ML systems with enterprise platforms Salesforce SAP ServiceNow Watson Assistant via REST APIs and webhooks

Implement real-time data streaming pipelines using Kafka and AWS Kinesis for live model inference

Collaboration & Delivery

Translate business requirements into ML/AI solution designs with clear success metrics and delivery milestones

Collaborate with data scientists to productionize research prototypes within agreed SLAs

Present model performance trade-offs and recommendations to technical and non-technical stakeholders

Participate in architecture reviews code reviews and sprint planning within Agile delivery teams

Mentor junior engineers and contribute to internal AI/ML best practice documentation

Stay current with state-of-the-art research and evaluate applicability to product use cases