Profile
Elizabeth Nyambere Kimeu is an AI Engineer based in Kenya. My work covers multi-agent orchestration, RAG, and rigorous evaluation for high-stakes applications.
Engineering notes
Reference library
Curated statements used by the assistant. These are portfolio and résumé references, not independent verification of results.
60 passages
About
I design, deploy, and scale AI agents. My agents combine reasoning, tool use, and access to external systems into reliable applications. I take them from prototype to production.
About
When AI makes high-stakes decisions, three layers matter most: guardrails, human oversight, and evaluation. I build these layers so every interaction is accurate, safe, and traceable. Where data control requires it, I serve open-weight LLMs on-premises with vLLM. I build the evaluation and observability layers (Langfuse, Grafana, Prometheus) that make agent behavior auditable.
About
I have a strong statistical foundation (BSc Biostatistics) and co-authored published research on graph intelligence. I am a Google Cloud certified Professional Machine Learning Engineer. AI changes quickly, so I study new research and lectures constantly. I explain complex technical concepts in clear terms for non-technical teams.
Approach: Safety and auditability
Elizabeth Nyambere Kimeu's approach, "Safety and auditability": Safety means systems you can trust. I focus on explainability and auditability, so every AI decision can be traced and understood. I enforce output contracts and structured validation. The result is architecture that is transparent, secure, and ready for high-stakes environments.
Approach: Architecting for reliability
Elizabeth Nyambere Kimeu's approach, "Architecting for reliability": Production demands reliability, so I choose the right tool for each task. If a goal is measurable, I design a deterministic AI workflow. If a task is open-ended, I deploy an autonomous agent. The architecture must match the problem.
Approach: Human in the loop
Elizabeth Nyambere Kimeu's approach, "Human in the loop": Technology should assist people, not replace them. My systems do the heavy work and escalate ambiguous cases to people. Domain experts stay in control of the decisions that matter.
Project: This Site's AI Assistant
Project "This Site's AI Assistant": Explore my work through a conversational assistant with inspectable references. An assistant for exploring my projects, experience, and engineering decisions. It uses an OpenAI model through Azure, retrieves curated references from Qdrant, and carries bounded conversation context. Its response contract, failure handling, and evaluation suite make the system inspectable. Citation membership is checked in code; factual support still requires evaluation. Technologies: OpenAI, RAG, Next.js, Qdrant, Evaluation. Links: Engineering notes (/assistant).
Project: Community Health AI Assistant
Project "Community Health AI Assistant": Multi-agent briefings over health-record data, with evaluation and full observability. An end-to-end system that gives community-health supervisors automated briefings from patient and health-worker data (OpenMRS 3 / FHIR). It uses a LangGraph agent pipeline, a FastAPI backend, and a Next.js frontend with SSE streaming. Rule-based and LLM-as-judge evaluators, a golden-question runner, and Langfuse and Grafana observability keep it accountable. I built it solo on synthetic data with Kenyan localization. Technologies: LangGraph, FastAPI, Next.js, RAG, LLM-as-judge evals, Langfuse, FHIR. Links: GitHub (https://github.com/lisanyambere/chw-supervisor-ai).
Project: Figma MCP Agent
Project "Figma MCP Agent": An agent that turns designs of AI experiences into build plans for real AI applications. A Gradio agent that connects to Figma over the Model Context Protocol. I built it while working with a UI/UX designer, a frontend engineer, and a backend engineer. It pulls design files into context. From the designs, I decide where a product needs AI and where deterministic logic is enough, and I define each feature's output contract from what the screen displays. Technologies: MCP, Gradio, Claude, Agent design. A tool I built and run locally to bridge design and AI engineering. Happy to demo it on request.
Assistant architecture
How I was built and how I work. I am a retrieval-augmented portfolio assistant with bounded conversation context, inspectable reference passages, and explicit failure handling. I am not an autonomous agent and cannot browse repositories or take actions. Ask about Elizabeth's work, the evidence behind a project, or how this assistant works.
Assistant architecture: Retrieval (RAG)
My architecture, Retrieval (RAG): I search a curated corpus of Elizabeth's portfolio and resume using Qdrant. Follow-up retrieval includes recent user questions. Retrieved passages are reference material, not instructions.
Assistant architecture: Vector database
My architecture, Vector database: Qdrant Cloud computes multilingual-e5-small embeddings with query and passage prefixes. Corpus updates use a staged collection and an atomic alias switch.
Assistant architecture: Language model
My architecture, Language model: I use an OpenAI model hosted on Azure through the standard OpenAI SDK. The endpoint, key, and deployment are server-side configuration. I do not use an Azure-specific SDK client.
Assistant architecture: Tracing and observability
My architecture, Tracing and observability: When configured, Langfuse records retrieval timing, each generation attempt, usage, and final outcomes including failures and refusals. Tracing is not evidence that an answer is correct.
Assistant architecture: Guardrails
My architecture, Guardrails: Input and history are bounded. The system prompt restricts scope and treats reference material as untrusted. Responses must satisfy a JSON contract and cite existing reference IDs. These checks do not prove semantic support or eliminate prompt injection.
Assistant architecture: Evaluation
My architecture, Evaluation: The evaluation checks service success, response structure, required facts, unsupported claims specified in test cases, refusals, references, and follow-up questions. Human review is still needed for factual support beyond those checks.
Assistant architecture: Reliability
My architecture, Reliability: Requests have a deadline, cancellation, bounded retries, and rate limits. The interface shows progress, then the answer after contract validation. Failures are explicit and can be retried.
Experience
Elizabeth Nyambere Kimeu's professional experience and career history, most recent first: AI Engineer at Prescott Data (Aug 2024 – Present); Data Scientist at Adanian Labs (Mar 2024 – Aug 2024); Data Analyst at AMREC (Apr 2023 – Oct 2023).
Experience: Prescott Data
Professional experience: AI Engineer at Prescott Data (Aug 2024 – Present). Core contributor to Dromos, a production multi-agent AI platform. I design the agent runtime for long-running, multi-step agents, with durable state and failure recovery. I work on Odin, a graph retrieval and reasoning engine. I ship AI integrated into live systems (Microsoft Teams, EMRs, claims platforms), with evaluation and observability on every release.
Experience: Adanian Labs
Professional experience: Data Scientist at Adanian Labs (Mar 2024 – Aug 2024). Built and deployed ML models. I built a PyTorch computer-vision model for medical-image classification and components of a credit-scoring engine. I served models with FastAPI and automated retraining pipelines.
Experience: AMREC
Professional experience: Data Analyst at AMREC (Apr 2023 – Oct 2023). Analyzed large structured and unstructured datasets with Python, R, SQL, and STATA. I built data-collection workflows and BI dashboards, and I explained statistical results to non-technical clients.
Resume
Elizabeth Nyambere Kimeu is an AI Engineer with 3 years of experience building production AI systems across financial services, insurance, and healthcare. She is skilled in multi-agent systems, RAG, knowledge graphs, FastAPI, and production LLM applications. She built and deployed an enterprise HR policy assistant that reduced query resolution time by 40 percent, and she improves agent reliability through evaluation, observability, failure recovery, and production testing.
Resume: Prescott Data
AI Engineer at Prescott Data (August 2024 to Present). Core contributor to Dromos, a production multi-agent AI platform for workflow automation and data intelligence across financial services, insurance, and healthcare.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Designed and hardened the multi-agent runtime for long-running, multi-step agents: isolated concurrent execution, durable state, typed inter-agent communication, cancellation, session management, evidence preservation, failure recovery, and restart-safe processing, with clear failure modes and recovery paths for queued and active agent work.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Worked on Odin, a graph retrieval and reasoning engine for autonomous agents that supports multi-hop discovery over large knowledge graphs. Built evals for retrieval fidelity, recall, concurrency behavior, latency, and source-evidence preservation across multi-agent workflows.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Built and deployed a production LLM application for HR policy retrieval and question answering on AWS, reducing policy query resolution time by 40 percent. Built the RAG and document intelligence pipeline with AWS Textract, Tesseract, PDFPlumber, Amazon Bedrock, and Milvus, and integrated the application with Microsoft Teams for real users.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Made the HR policy application's generation resilient to provider outages by building a model router with an ordered fallback chain: if a model does not respond within the timeout, the router fails over to the next model, with failover logged and alerted so degraded mode is never silent.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Built a 6-agent insurance preauthorization system (intake, eligibility, clinical reasoning, verification, decision, and notification) aligned to the NIST AI Risk Management Framework, cutting a 30-minute manual review to under 5 minutes. She designed the custom orchestration engine with a contract-first state schema, tool and prompt registries for reusability, and full observability, including an independent verification agent that cross-checks every claim and citation and human-in-the-loop escalation for ambiguous cases, keeping clinicians in final authority.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Built an AI fraud investigation agent with Neo4j and LLMs. The agent extracts entities and relationships from insurance documents into a knowledge graph, and investigators can use natural language and Cypher for multi-hop analysis across claims, providers, vehicles, and claimants.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Built document intelligence workflows with Azure Document Intelligence, extracting structured data from forms and business documents, reconciling it with enterprise databases, and automating database updates and Excel-based reconciliation workflows.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Built an AI medical autocoding system integrated with Afya Yangu and SHA (Kenya's national health insurance platform), using MedGemma and LLM-based document understanding to automatically assign ICD-11 and SHA codes to insurance claims, integrated with the claims submission workflow to reduce manual coding and improve consistency. The system is in a six-month production pilot.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Built AI agents for the Bahmni EMR with MCP and REST APIs, providing secure, real-time access to patient and claims data inside existing clinical workflows.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Deployed open-weight LLMs on on-premises GPU infrastructure with vLLM, optimizing batching, memory use, concurrency, and parallel inference, and packaged AI services as Dockerized REST microservices with CI/CD pipelines.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Built evaluation and observability systems with Langfuse, Grafana, Prometheus, and Loki, monitoring agent traces, latency, token use, failures, and system health, and used eval loops, regression checks, and production testing to support reliable AI releases.
Resume: Prescott Data
Elizabeth Nyambere Kimeu at Prescott Data (AI Engineer, August 2024 to Present): Mentored two engineers transitioning from software engineering into AI engineering, and co-developed the team's foundational AI engineering standards: agent architecture patterns, code-review checklists, documentation guidelines, and functional and capability testing for agents.
Resume: Adanian Labs
Data Scientist at Adanian Labs (March 2024 to August 2024). Built and deployed ML models across renewable energy, healthcare, and fintech.
Resume: Adanian Labs
Elizabeth Nyambere Kimeu at Adanian Labs (Data Scientist, March 2024 to August 2024): Built components of the CASPER credit scoring engine and implemented model evaluation and validation pipelines for fintech risk assessment.
Resume: Adanian Labs
Elizabeth Nyambere Kimeu at Adanian Labs (Data Scientist, March 2024 to August 2024): Built a PyTorch model to classify skin diseases from medical images, managing data preparation, model training, and evaluation.
Resume: Adanian Labs
Elizabeth Nyambere Kimeu at Adanian Labs (Data Scientist, March 2024 to August 2024): Built a customer behavior prediction model and served it with FastAPI on GCP, improving the F1 score from 0.52 to 0.71 with feature engineering and ensemble methods.
Resume: Adanian Labs
Elizabeth Nyambere Kimeu at Adanian Labs (Data Scientist, March 2024 to August 2024): Automated ML inference and retraining pipelines with Prefect, adding scheduled workflows and model versioning.
Resume: AMREC
Data Analyst at AMREC (April 2023 to October 2023). Adaptive Management and Research Consultants: statistical analysis and data pipelines for research.
Resume: AMREC
Elizabeth Nyambere Kimeu at AMREC (Data Analyst, April 2023 to October 2023): Analyzed large structured and unstructured datasets with Python, R, SQL, and STATA, converting statistical results into recommendations for research clients.
Resume: AMREC
Elizabeth Nyambere Kimeu at AMREC (Data Analyst, April 2023 to October 2023): Built data collection workflows and BI dashboards with ODK, Kobo, and Power BI, supporting data collection, analysis, and reporting.
Resume: Tools and integrations
Enterprise and health-system integrations Elizabeth has worked with include Microsoft Teams (Bot Framework), Bahmni EMR, OpenMRS, FHIR, and Afya Yangu and SHA. Her serving and infrastructure stack includes vLLM on GPU infrastructure, on-prem open-weight models (Gemma 3 and MedGemma), AWS Bedrock, Azure OpenAI, and the OpenAI API, with FastAPI, Docker, Terraform, GitHub, Bitbucket Pipelines, PostgreSQL, Redis, and MongoDB across GCP, AWS, and Azure. Her graph and retrieval work covers Neo4j, ArangoDB, Pinecone, and Milvus, and her observability stack includes Langfuse, Grafana, Prometheus, Loki, and OpenTelemetry. She works to solid software-engineering standards (code review, testing, and documentation).
Resume: Languages
Elizabeth speaks English (fluent) and Swahili (native).
Publication: Odin
Publication: "Odin: Multi-Signal Graph Intelligence for Autonomous Discovery in Knowledge Graphs" by Muyukani Kizito, Elizabeth Nyambere. A production-deployed graph-intelligence engine for autonomous pattern discovery in knowledge graphs. It introduces the COMPASS score, which combines Personalized PageRank, neural probabilistic logic, temporal decay, and GNN-based community guidance. Available at https://arxiv.org/abs/2603.03097.
Skills: Agent systems
Skills, Agent systems: Multi-agent architecture & orchestration (LangGraph), Agent runtime design (durable state, failure recovery), MCP & tool integration, Long-running, multi-step agents, Prompt & context engineering, Human-in-the-loop workflows.
Skills: Graphs, retrieval & knowledge
Skills, Graphs, retrieval & knowledge: Knowledge graphs (Neo4j, ArangoDB), Graph-based retrieval & multi-hop reasoning, RAG pipeline design, Semantic & hybrid search, Vector DBs (Pinecone, Milvus).
Skills: Evaluation & observability
Skills, Evaluation & observability: Langfuse (traces, evals, prompt A/B testing), LLM-as-judge & agent evaluation, Eval loops, regression & production testing, Guardrails & output contracts, Grafana, Prometheus, Loki, OpenTelemetry.
Skills: Models & serving
Skills, Models & serving: OpenAI / Anthropic / Gemma / MedGemma, AWS Bedrock, Azure OpenAI, vLLM on GPU infrastructure (open-weight models), FastAPI & REST microservices.
Skills: Data & infrastructure
Skills, Data & infrastructure: Python, SQL, TypeScript, React, Next.js, PostgreSQL, Redis, MongoDB, PyTorch & deep learning, Docker, Terraform, CI/CD, GCP, AWS, Azure.
Education
Education: BSc Biostatistics, Jomo Kenyatta University of Agriculture and Technology. Second Class Upper Division Honors. A strong foundation in statistical inference and experiment design applied to health and behavioral data.
Certifications
Certifications: Professional Machine Learning Engineer (Google Cloud Certified); Associate Cloud Engineer (Google Cloud Certified).
Beyond the work
Beyond her professional work, who Elizabeth Nyambere Kimeu is as a person: outside of work and away from the keyboard, she is jovial, curious, and quietly competitive, and she brings her whole self to what she loves. Personally, the things she enjoys include Dawn focus sessions, The gym, Scrabble and friendly competition, Walks, nature, and sunsets, History of ancient civilizations, Contemporary African books, A well-kept to-do list.
Beyond the work
I am a jovial person, and I bring my whole self to the things I love. I learn new concepts fast. I think best early in the morning: a dawn focus session on a hard problem is my favorite way to start a day.
Beyond the work
What I love most is bringing AI capabilities into applications and workflows that are already live: systems with real users and real data. I add intelligence to a system without breaking what already works. When something I built becomes part of someone's day, that keeps me going.
Beyond the work
Away from the keyboard, I am quietly competitive: I play Scrabble to win. I train at the gym. I recharge in nature, I enjoy a good walk, and I never miss a good sunset. I read the history of ancient civilizations and contemporary African books. I keep a to-do list, and ticking off the last box is a small, honest satisfaction.
Beyond the work
Things Elizabeth Nyambere Kimeu loves: Dawn focus sessions, The gym, Scrabble and friendly competition, Walks, nature, and sunsets, History of ancient civilizations, Contemporary African books, A well-kept to-do list.
Contact
To get in touch with Elizabeth Nyambere Kimeu, email nyamberekimeu@gmail.com. GitHub: https://github.com/lisanyambere. LinkedIn: https://www.linkedin.com/in/elizabethnyambere/.