Work Schedule
Standard (Mon-Fri)
Environmental Conditions
Office
Job Description
Thermo Fisher Scientific Inc. is the world leader in serving science, with annual revenue exceeding $40 billion and extensive investment in R&D. Our Mission is to enable our customers to make the world healthier, cleaner and safer. The customers we serve fall within pharmaceutical and biotech companies, hospitals and clinical diagnostic labs, universities, research institutions and government agencies. Our innovations drive scientific breakthroughs, from groundbreaking research to routine testing and real-world applications.
How will you make an impact?
You will play a pivotal role in designing and delivering reliable, robust AI applications, algorithms, and frameworks that elevate the quality and performance of our product offerings.
You will collaborate with and learn from a dedicated team of algorithm and software developers, revolutionizing healthcare through low-cost and high efficiency diagnostic systems.
What will you do?
- Develop, prototype, and deploy AI applications—including chatbots, copilots, and autonomous agents—optimized for safety, performance, and cost-efficiency.
- Build and optimize retrieval-augmented generation (RAG) pipelines (indexing, chunking, embedding, reranking, grounding).
- Implement parameter-efficient methods (LoRA/QLoRA/PEFT) and model optimization workflows.
- Prepare, clean and facilitate the annotation of datasets for LLM model training and improvement.
- Apply context engineering (prompt design, tool calling, memory, compression, window strategy).
- Establish evaluation: golden sets, RAG/grounding scores, toxicity, A/B tests, latency & cost measurements.
- Integrate tools via Model Context Protocol (MCP) and other agent frameworks.
- Design and implement robust, multi-turn conversational agents with clear handoff strategies—such as transitioning to human support or fallback responses—alongside integrated safety guardrails and seamless deployment across web, mobile, and internal platforms.
- Operate in production: tracing, prompt/version lineage, drift detection, incident response, SLOs.
- Collaborate with cross-functional teams, including software, hardware, and data science, to ensure algorithms meet product requirements and are well-integrated into production systems.
- Contribute to code quality, documentation, and reproducibility, ensuring algorithm reliability and transparency.
- Stay up-to-date on new technologies and industry standards to continuously enhance development and evaluation methodologies.
How will you get here?
Education
Master’s degree in Computer Sciences, Mathematics, Bioinformatics or a related field; PhD or equivalent experience is highly preferred.
Experience and skills Required
- Ideal candidates will have a minimum of 2 years of experience in AI development and evaluation. Recent graduates who have conducted relevant research are also encouraged to submit their applications.
- Hands-on experience with major LLMs/APIs or frameworks (OpenAI, LangChain or Anthropic, Hugging Face etc).
- Strong prompt & context engineering.
- Strong programming skills in Python and familiarity with Java or JavaScript for integration tasks.
- Understanding of data preprocessing, distributed training and model evaluation.
- Experience fine-tuning LLMs (i.e. GPT, LLaMA, Mistral) using PEFT techniques like LoRA/QLoRA.
- Strong communication skills, and the ability to present work to both technical specialists and non-experts.
- Ability to work both independently and collaboratively.
- Highly motivated and fast learner: able to proactively scope ambiguous tasks, set clear goals, ship small iterations, seek feedback, and reliably follow through.
Preferred
- Familiarity with RAG frameworks with vector search (e.g., FAISS/pgvector/Pinecone), hybrid search, rerankers, citation/grounding.
- Experience with MLOps tools (MLflow, Kubeflow, Docker, CI/CD) and cloud services.
- Hands-on experience in building agents or chatbots.
- Evaluation: RAGAS/G-Eval or similar; offline/online metrics and A/B frameworks.
- Data systems: SQL/NoSQL, message queues, object storage; schema design for documents & metadata.
- Understanding of AI safety principles, hallucination mitigation, and factuality evaluation techniques.
- Hands-on experience with AWS SageMaker, Bedrock, and Step Functions, along with other relevant AWS services, to build, deploy, and orchestrate AI agents in scalable, production-grade workflows.
- Experience in the biotechnology industry is a plus.