Bayesoft · 1 day ago
AI Engineer
Illinois, United States
Full-time
Onsite
Mid, Senior Level
3+ years expBayesoft is an innovative early-stage startup revolutionizing drug development by integrating Bayesian statistics with cutting-edge AI technologies. They are seeking a hands-on AI Engineer to design, build, and deploy intelligent features within their applications, collaborating closely with statisticians and Data Scientists. The role involves scoping, prototyping, and implementing AI products from model design to integration.
Research
No H1B
Responsibilities
Design and implement domain-aware AI agents capable of multi-step reasoning, constraint satisfaction, and decision support for clinical trial design and development workflows
Architect and implement agentic AI systems using explicit state machines, including planning, tool selection, memory management, error recovery, and deterministic replay for reproducibility and auditability
Build production-grade AI services and APIs (FastAPI or equivalent) that expose agentic workflows to downstream applications, ensuring scalability, observability, and versioned behavior
Design evaluation and validation pipelines for agent behavior, including correctness checks, numerical validation (e.g., statistical outputs) and failure mode analysis
Lead technical exploration of advanced AI architectures including RAG, GraphRAG, knowledge-graph-augmented reasoning, and hybrid symbolic–LLM systems in regulated domains
Work with structured and unstructured data stored in AWS RDS, SQL Server, and other data sources
Qualification
Required
Demonstrated experience building complex AI or ML-driven systems end-to-end. Preferred 3–5 years of experience as an AI Engineer or Software Engineer
Experience implementing multi-step or multi-agent AI systems with explicit control flow, tool orchestration, and intermediate state management (not prompt-only chatbots)
Strong proficiency in Python, including experience with agent orchestration frameworks (e.g., LangGraph), LLM integration libraries (LangChain or equivalent), and ML frameworks such as PyTorch, with an emphasis on building reliable, modular AI systems
Experience with LLM integration and prompt engineering (e.g., OpenAI, Anthropic Claude, Gemini, etc.)
Experience deploying chatbots, retrieval-augmented generation (RAG), or embedding-based search
Understanding of MLOps concepts such as versioning, CI/CD, and monitoring
Experience translating complex domain logic into AI-assisted reasoning pipelines, including symbolic logic, statistical computation, and LLM-based decision support
Deep experience with tool-augmented LLM systems, including structured tool calling, schema-constrained outputs, and recovery from tool or execution failures
Knowledge of API integration and orchestration frameworks (FastAPI, Flask, or Streamlit)
Preferred
Background or strong interest in statistics, biostatistics, or clinical trial methodology, including familiarity with randomized controlled trials
Experience with or strong interest in LLM post-training techniques, such as supervised fine-tuning, preference optimization, or reinforcement learning–based approaches for shaping model behavior
Familiarity with agent learning paradigms, including feedback-driven improvement, reward modeling, or policy optimization in tool-augmented or multi-step reasoning systems
Experience in applying AI complex domains with large numbers of entities and relationships
Proven track record in building AI applications for end-users
Understanding of responsible AI principles and data governance best practices
Experience with SQL and data modeling using AWS RDS or SQL Server
Prior knowledge on Ontology/knowledge graph, Graph Database (Neo4j, ArangoDB, RDF/SPARQL; entity linking & ID mapping (RxNorm, MedDRA/CTCAE))
Company
Bayesoft
Bayesoft is an innovative early-stage startup revolutionizing drug development by integrating Bayesian statistics, advanced Bayesian clinical trial designs, and cutting-edge AI technologies.
2-10 employees
Funding
Current Stage
Early StageCompany data provided by crunchbase