AI Engineer

About AjentikAjentik AI Pte. Ltd. is a Singapore healthtech company incubated at NUSX (formerly NUS Enterprise). We build Elderwise, an AI-powered Comprehensive Geriatric Assessment (CGA) platform that turns caregiver observations into structured, EMR-ready clinical summaries. Caregivers speak with our voice assistant in the browser, with no app to install. Clinicians receive a structured summary to review and file, and families receive a plain-language report. We are in pilot discussions with geriatric teams across Singapore's public healthcare clusters and the community care sector.About the RoleAs an AI Engineer, you will build the conversational and extraction pipelines at the heart of Elderwise: the voice conversation with a caregiver, the structured clinical fields it produces, and the summary a clinician reviews and signs. You will ship to real users in a setting where accuracy, privacy and clinician trust matter more than demo polish.The role suits an engineer with one to four years of experience who has shipped software to production and wants to go deep on applied LLM engineering. You will work directly with the founders, our engineers and our Clinical Lead, with experienced engineers reviewing your designs and code. Your scope will grow as you show you can own it.Key ResponsibilitiesBuild and improve LLM pipelines end to end: conversation design, prompting, structured outputs, schema validation, retries, logging and storageTurn caregiver conversations into validated clinical fields, each traceable to what was said and marked as stated, inferred or not askedIntegrate speech-to-text and text-to-speech into low-latency voice conversations, and catch transcription errors before they reach a clinical documentGrow our evaluation suite: golden sets, synthetic caregiver personas, adversarial cases and regression checks that run in CI on every prompt or model changeTrack quality, cost and latency per release, and make model and prompt changes on evidence rather than intuitionPush the boundaries of what our AI can do: prototype next-level capabilities such as emotion and stress detection from voice and conversation, measure them rigorously with our Clinical Lead, and take what works from experiment into the productBuild guardrails that hold every time: no diagnosis or medical advice, safe escalation paths and defences against prompt injectionProduce interoperable outputs (structured JSON, PDF reports and HL7 FHIR R4 resources) that fit how hospitals file recordsHandle personal and health data responsibly under PDPA and our ISO 27001 and ISO 42001 programmes, from data minimisation to redacted logsWork spec-first with AI coding agents: write the spec and tests, let the agent implement, and own the resultJoin clinician and caregiver sessions to see how the product is used, and turn what you learn into engineering workRequirements1 to 4 years of professional software engineering experience, with strong PythonAt least one LLM-powered feature or product you built that real users relied on, not only a demo, and the ability to explain what broke and how you fixed itWorking knowledge of structured outputs, prompt design and the ways LLMs fail, including hallucination, non-determinism and behaviour changes between model versionsA habit of measuring: you write tests and evaluations, not only promptsSolid fundamentals in APIs, SQL, Git, automated testing and CIClear written communication in English, and honesty about what you know and what you don'tCare for privacy and safety when working with sensitive personal dataNice to HaveExperience with voice or speech systems: streaming STT and TTS, turn-taking and realtime audioFastAPI and Pydantic, PostgreSQL with pgvector, Supabase, Docker and KubernetesRetrieval (hybrid search, reranking) and agent tool use, including the Model Context Protocol (MCP)Serving open-weight models (for example with vLLM) or running small models on deviceResearch or prototyping in affective computing, speech emotion recognition or multimodal modelsHealthcare data standards such as HL7 FHIR, or experience in healthtech or another regulated domainConversational Mandarin, Malay or Tamil, to help evaluate multilingual caregiver conversationsA degree in computer science, engineering or a related field, or equivalent practical experienceWhat We OfferCompetitive salary with equity participationReal ownership of production AI systems from your first monthRoom to explore new AI capabilities and take the best ideas into productionMentorship from experienced engineers, our Clinical Lead and the foundersDirect contact with the clinicians and caregivers who use what you buildAccess to frontier and open-weight models and modern AI development toolingHealth benefits and flexible work arrangementsLearning budget for courses and conferencesA clear path toward Senior AI Engineer as the engineering team growsYour ImpactEvery field our pipeline captures correctly is one less question a clinician has to ask again, and every hour we save goes back to patient care. Your work will decide whether clinicians can trust AI with the information that shapes care for older adults, in Singapore and beyond.Ajentik is an equal opportunity employer. We welcome applicants of all backgrounds and assess every candidate on merit.