Senior Data Scientist - Conversational AI & Customer Intelligence

About the role:

Ringkas operates an agentic financial services platform that powers AI-driven customer conversations at scale for financial institutions. Our AI agent manages the full inbound customer journey — from a targeted campaign outreach through to a completed financial product application — handling conversion end-to-end without additional operational headcount for our partners.

We are looking for a Senior Data Scientist to deepen the intelligence layer that drives our agent’s effectiveness. Our AI engineering team runs and evolves the agent itself. Your role is to be the analytical brain behind it: sharpening the frameworks for each customer segment, extracting richer insight from growing volumes of conversation data, and continuously improving how the agent reads and responds to each type of customer — to maximise engagement and conversion outcomes for our partners.

This is a research and insight role with direct product impact. The frameworks and analytical outputs you produce are operationalised by the engineering team into live agent behaviour. As our conversational dataset deepens across product types, customer segments, and markets, the insight layer is already one of the most defensible assets in the platform — and this role is central to realising its full value.

WHAT YOU WILL DO

Conversational Analytics (Primary)

Conversational Analytics (Primary) Design and own the conversation analysis pipeline: per-turn sentiment scoring, intent classification, objection detection, and session-level funnel analysis Build and maintain the intent taxonomy for Indonesian financial conversations — iterating the classification schema as new customer behaviours emerge Analyse session-level patterns to identify where customers disengage, what objection types cluster at each funnel stage, and which conversational sequences most strongly predict application completion Apply topic modelling (e.g. BERTopic) to surface emerging customer concerns and unmet product needs across campaign cohorts Define the analytical framework for each customer segment — what good engagement looks like, which intent sequences signal genuine purchase readiness, which objection patterns indicate a segment is poorly matched to the product being offered Develop the feedback loop: as session outcomes resolve (converted, dropped, deferred), systematically update your frameworks and surface findings to the AI engineering team for implementation in the agent. 2. Modelling & Inference Build propensity models using session-level behavioral features (response latency, session depth, sentiment trajectory, intent sequence) as inputs alongside customer data Implement Markov-style intent transition models to estimate conversion probability from observable conversation state Design and analyse A/B experiments on message variants, offer framing, and conversation flow — with rigorous attribution back to conversion delta 3. LLM Application & Evaluation Design and iterate structured prompts that reliably extract sentiment, intent, objections, and topics from Indonesian financial conversations — returning clean, parseable JSON outputs Build the pipeline that routes conversation data through LLM APIs (Anthropic, OpenAI, or equivalent) in efficient batches, with validation, error handling, and cost tracking Develop and maintain a labeled evaluation set to measure prompt performance systematically — treating prompt iteration the way a traditional ML engineer would treat model retraining Know when and how to substitute a smaller or more efficient model for high-volume tasks once performance baselines are established — our philosophy is best-performing system first, optimise cost later 4. Conversion Intelligence Combine conversational signals with product uptake data — who applied, which product, at which point in the conversation — to build a continuously improving picture of what drives customers to convert Identify which conversational patterns, customer profiles, and segment configurations produce the strongest conversion outcomes, and translate these into concrete handling recommendations for the engineering team As the platform expands across product types and markets, develop a framework for how segment intelligence transfers — what is universal across financial products and what needs to be rebuilt from scratch for each new context Work closely with the AI engineering team to communicate findings in a form they can implement: structured schemas, decision rules, and prioritised insight rather than open-ended reports 5. Insight & Communication Produce campaign-level insight reports for bank partners: dominant sentiment trajectories, top objection themes, funnel break points, and conversion drivers Translate analytical findings into concrete campaign recommendations — message adjustments, re-contact timing, segment refinements Present findings clearly to non-technical stakeholders including bank business and product teams

WHAT WE ARE LOOKING FOR

Required

Strong academic background in statistics, econometrics, mathematics, computer science, or a related quantitative discipline 3–6 years of hands-on data science or applied AI experience, with demonstrable work in text analytics, conversational data, or LLM-based systems Fluency with LLM APIs (Anthropic, OpenAI, or equivalent): structured prompting, JSON output parsing, batch processing, and prompt evaluation — this is your primary technical tool Solid statistical foundation — regression, classification, survival analysis, experiment design. Econometrics or quantitative social science backgrounds are a genuine advantage here Direct experience working with text data in Bahasa Indonesia — you understand informal register, code-switching, and how Indonesians actually write in chat Ability to work with raw, messy data — you will be building pipelines from conversation logs, not receiving clean feature tables Clear written and verbal communication in both Bahasa Indonesia and English — you will write insight reports that a bank's business team needs to act on

Strong Advantage

Experience with conversational data specifically: chat logs, call transcripts, or any sequential text where the unit of analysis is a session rather than a document Familiarity with financial products and how customers talk about them — mortgages, savings, loans, insurance — across any market. Indonesian financial product context is a plus given our current focus, but not a hard requirement Experience in a fintech, financial services, or any high-stakes conversion environment (e-commerce, insurance, lending) where understanding why customers do or do not complete a transaction was the central analytical problem Familiarity with model cost optimisation — understanding the trade-offs between frontier and smaller models, and practical experience making that switch in a real system Experience building lightweight dashboards or reporting outputs (Streamlit, Metabase, or similar)

HOW WE WORK

Small founding team — your work has direct product impact from week one You will own the analytical layer end-to-end: no handoffs to a separate data engineering team at this stage English and Bahasa Indonesia are both working languages We move fast and iterate based on real campaign data — you will see the results of your models playing out in live customer conversations

WHAT WE OFFER Competitive compensation Hybrid working arrangement Multicultural, high-performance and fun environment

HOW TO APPLY Apply by clicking the Apply button on this posting. The form will ask for: Your email address Your CV (PDF upload) A short written answer: why are you the right candidate for this role? We are looking for a specific, considered response — not a generic cover letter. Candidates who do not complete this section will not be reviewed. If your profile is selected for assessment, you will receive an email within 2 working days. If you do not hear from us within that window, your application was not taken forward for this round.