Data Scientist

Data Scientist

Location : Houston

Contract Type : 6-Month Ongoing Contract (potential to hire Fulltime)

Overview

• Design & deliver GenAI solutions: Architect and implement LLM/LVM applications (text and, where applicable, vision) including prompt strategies, guardrails, evaluation metrics, cost/latency optimization, and production rollout.
• Build robust RAG systems: Stand up end‑to‑end RAG pipelines (ingestion → chunking → embedding → retrieval → synthesis), with observability, feedback loops, and AB testing for groundedness and hallucination control.
• Fine‑tune foundation models: Select, adapt, and fine‑tune open and hosted models for domain‑specific tasks using efficient techniques (LoRA/QLoRA, PEFT, parameter‑efficient adapters), and manage evaluation datasets.
• Develop agentic workflows: Implement multi‑step, tool‑using agents for real‑time, context‑aware operations; orchestrate planning, memory, and tool calling with safe‑execution policies.
• ML/LLM engineering: Build high‑quality Python code, reusable libraries, and APIs; implement CI/CD, testing, experiment tracking, and model/version governance.
• Data & platforms: Partner with data engineering to operationalize pipelines on enterprise platforms (e.g., Databricks/Snowflake) and integrate with cloud AI/ML services.

Required qualifications

Mathematics & foundations

• Statistics , Multivariate Calculus , Linear Algebra , Optimization (you can explain choices and trade‑offs in model behavior based on these principles).

Programming

• Advanced Python (clean architecture, typing, packaging, testing; performance profiling and async where appropriate).

Python ecosystem

• Data Handling/Visualization: pandas, NumPy, Seaborn/Matplotlib, pyspark
• Machine Learning: scikit‑learn, XGBoost, LightGBM
• Deep Learning:TensorFlow or PyTorch
• Generative AI:LangChain , LlamaIndex , Haystack , Hugging Face transformers

Generative AI expertise

• Large Language/Vision Models (LLM/LVM): Hands‑on with multiple providers/models (e.g., Gemini , GPT , Claude , Llama ) and their APIs.
• Model fine‑tuning: Proven experience fine‑tuning foundation models for domain tasks; evaluation design and data curation included.
• Retrieval‑Augmented Generation (RAG): Ability to design and implement robust RAG systems for real‑time, context‑aware applications.
• Prompt engineering & agentic workflows: Advanced prompt design (system/task/reflection patterns) and building multi‑step AI agents .
• Vector databases/search & embeddings: Practical experience with vector indexing, similarity search, and embedding selection/management.

Software craftsmanship & platforms

• Version Control: Git (GitHub/GitLab/Bitbucket)
• Databases:SQL and NoSQL (e.g., MongoDB , Cassandra )
• Cloud: Hands‑on with one or more of AWS , Azure , GCP , specifically AI/ML & data services (e.g., AWS SageMaker , Azure Machine Learning , Google Vertex AI )
• Enterprise data engineering platforms:Databricks , Snowflake
• Azure AI Foundry:Hands‑on experience required