Singapore · Comp Sci & Design · SUTD 2026

meghapusti

Currently
AI / ML Engineer
Scroll to explore

I build systems that think. From production RAG pipelines used daily by research scientists, to LLM-powered compliance platforms, to geopolitical intelligence tools trained on a decade of global data — I close the gap between what AI can do and what people actually need.

Merit Scholar · Minors: AI & Digital Humanities · Specialisations: FinTech & Data Analytics

20+
Projects
P&G Stints
100+
Collaborators Served
Curiosity
AI / ML EngineeringRAG PipelinesLLM IntegrationFull-Stack SystemsNLP ResearchVector DatabasesData EngineeringGeopolitical IntelligenceFinancial TechnologyVR DevelopmentDatabase SystemsDigital Humanities AI / ML EngineeringRAG PipelinesLLM IntegrationFull-Stack SystemsNLP ResearchVector DatabasesData EngineeringGeopolitical IntelligenceFinancial TechnologyVR DevelopmentDatabase SystemsDigital Humanities
01 —

About Me

I'm a final-year Computer Science & Design student at SUTD, specialising in AI systems and data engineering. My core interest is the space where machine learning meets production infrastructure — building pipelines that don't just work in notebooks, but in the real hands of real people.

Two internships at Procter & Gamble's Singapore Innovation Center shaped how I think about ML work: I benchmarked retrieval architectures to identify the optimal RAG setup for internal R&D, built Text-to-SQL systems enabling 50+ non-technical scientists to query complex research databases, and developed multimodal databases linking visual and sensory attributes with generative AI. Useful, actually deployed, measurably impactful.

My projects reflect the same range: a geopolitical intelligence platform with production LightGBM classifiers trained on 11 years of GDELT data; an enterprise lead management system with a security-first AI chatbot; an NLP study proving translation pipelines destroy emotional meaning in Singlish. I also hold a Minor in Digital Humanities — and yes, I have used Voyant Tools corpus analysis to validate a literary theory argument about Plato and Beckett.

DegreeBE Computer Science & Design, SUTD
MinorsArtificial Intelligence · Digital Humanities
TracksFinancial Technology · Data Analytics
AwardSUTD Undergraduate Merit Scholarship
BasedSingapore · Open to Relocation
Technical Stack
RAG & LLMs
FAISSPineconeOpenAI APIAzure OpenAISentence-TransformersLangChainRAG PipelinesRe-ranking
ML / AI
PyTorchTensorFlowHuggingFaceLightGBMCatBoostscikit-learnSHAPONNXOptuna
NLP
BERT / DistilBERTT5spaCyNLTKSBERTFinBERTDeBERTaBiLSTM
Full-Stack & APIs
ReactNext.jsNode.jsFastAPIFlaskExpressStreamlitREST / OpenAPI
Data & Databases
PostgreSQLTimescaleDBMongoDBSQLiteFirebaseSequelizeSQLAlchemyETL Pipelines
Cloud & DevOps
GCP (Cloud Run)AWSDockerRailwayGitVercel
Languages
PythonJavaCSQLJavaScriptC#Lucid HDL
Tools & Other
Unity 6FPGA (Xilinx)Power BITableauFigmaBloomberg TerminalVoyant ToolsWireshark
02 —

Work Experience

May – Sep 2025
Procter & Gamble
Singapore Innovation Center
Internship II · AI/ML
RAG Vector Search Re-ranking Benchmarking Production
Details
Associate Scientist Intern, Data Science & Modeling RAG · Retrieval · Production
  • Identified the optimal production RAG architecture by benchmarking 5+ retrieval models, vector database configurations, and re-ranking strategies — improving internal R&D knowledge retrieval accuracy across the team.
  • Designed and deployed end-to-end RAG pipelines integrating vector databases, embedding models, and re-ranking layers; meaningfully reduced the time scientists spent searching hundreds of internal research documents.
  • Established reproducible evaluation benchmarks from large-scale historical consumer research datasets, enabling consistent and auditable model selection across retrieval experiments.
  • Iteratively reduced retrieval error rates through empirical experimentation conducted in close collaboration with domain experts, improving system relevance scores across successive evaluation cycles.
Aug 2024 – Jan 2025
Procter & Gamble
Singapore Innovation Center
Internship I · GenAI
GenAI Text-to-SQL Multimodal DB NLP 50+ Users
Details
Associate Scientist Intern, Data Science & Modeling GenAI · NLP · Text-to-SQL
  • Built a multimodal research database linking visual and sensory product attributes using generative AI, enabling large-scale analysis of consumer perception patterns across thousands of samples.
  • Surfaced statistically significant relationships between visual cues and sensory descriptors by analysing multi-year consumer research corpora, providing actionable R&D insights to stakeholders.
  • Implemented a Text-to-SQL system enabling natural language querying across complex multi-table research databases — democratising data access for 50+ non-technical researchers with zero SQL knowledge required.
  • Validated all model outputs with domain experts before production handoff, ensuring full reliability and alignment with established research findings.
Sep – Dec 2025
Outplat
Australia · Remote
AI Engineer · Full-Stack
GRC Compliance RAG LLM Pipelines Full-Stack Cloud
Details
AI Engineer & Full-Stack Developer Compliance AI · GRC · RAG
  • Contributed to an AI-powered GRC compliance platform unifying data from multiple cybersecurity tools into a single risk dashboard, reducing manual aggregation overhead for enterprise clients.
  • Built RAG-based LLM pipelines generating contextualised, policy-specific compliance recommendations from unstructured regulatory documents.
  • Developed modular backend and frontend components on a cloud-hosted stack, enabling repeatable deployment and scalable enterprise client onboarding.
Mar 2023 – Present
Young Engineers SG
Singapore
Instructor
Python Robotics 100+ Students Mentorship
Details
Programming & Robotics Instructor Teaching · Mentorship
  • Improved programming proficiency across 100+ students aged 3–16 through structured project-based instruction in Python, robotics, and introductory machine learning.
  • Mentored students for competitions, developing computational thinking and systematic debugging mindsets applicable to real engineering problems.
03 —

Selected Projects

04 —

Get in Touch

Let's build
something
worth building.

I'm actively looking for full-time roles in AI/ML engineering, data science, or full-stack development. Singapore-based but open to relocation globally. If you're working on something interesting — reach out.

Let's Connect! →