Niranjana

Key Skills

Software

Other

Top Subject Matter

Biomedical ML (retinal imaging, genomics, ECG) and psychological well-being analytics

Banking chatbot safety and evaluation

LLM workflow automation and evaluation for software lifecycle artifacts

Top Data Types

Text

Audio

Image

Top Task Types

Prompt + Response Writing (SFT)

Function Calling

Classification

Entity (NER) Classification

Freelancer Overview

Research Assistant (supervised by Prof. Imran Razzak) at UNSW Sydney. Brings 4+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other. Education includes Bachelor of Computer Science, Honours, UNSW Sydney (2025). AI-training focus includes data types such as Text and Audio and labeling workflows including Evaluation, Rating, and Prompt + Response Writing (SFT).

Labeling Experience

Graduate AI Engineer at Westpac Banking Corporation

OtherTextFunction Calling

Developed agentic AI workflows to automate software-lifecycle artifacts using LLM-generated outputs and ongoing LLM evaluation. Created LLM-driven user stories and test cases and generated analytics and reporting to assess performance over time. Implemented systems for identifying redundant work via semantic similarity between text artifacts. • LLM-driven generation of user stories, analytics, and reporting • LLM evaluation to monitor quality during automated workflow execution • Semantic similarity and recommendation system using embeddings for redundancy flagging • Designed workflows for software lifecycle automation with LLM-driven components

2026 - Present

Research Assistant (supervised by Prof. Imran Razzak) at UNSW Sydney

OtherText

Conducted ML, NLP, and computer-vision modeling work on biomedical datasets to support disease prediction and biomarker discovery from clinical signals. Transformed multimodal inputs into model-ready representations for supervised learning objectives. Worked on study development and model validation steps associated with biomedical inference tasks. • Disease prediction and biomarker discovery using retinal imaging, genomics, and ECG • Statistical and machine-learning methods for psychological well-being drivers • Applied NLP/ML pipelines to derive predictive features from biomedical text and signals • Co-authored research outputs tied to model evaluation and inference performance

2024 - 2025

Data Science Intern - Commonwealth Bank of Australia

AudioClassificationEntity (NER) Classification

Develop and prototype chatbot features for the Bankwest Broker model, including user feedback mechanisms and text-to-audio capabilities. Build and test chatbot guardrails and adversarial prompts to improve robustness and reduce invalid prompt handling. Apply product-minded engineering practices together with NLP/AI tooling from a banking context. • Implement chatbot feedback and text-to-audio feature ideas • Design guardrails for safer, more reliable conversational behavior • Create adversarial test prompts during a hackathon to evaluate failure modes • Improve invalid-prompt flagging accuracy using iterative experimentation

2023 - 2024

Data Science Intern at Commonwealth Bank of Australia / Bankwest Broker model

OtherTextPrompt + Response Writing (SFT)

Built and stress-tested chatbot functionality by designing guardrails and adversarial prompting strategies to improve invalid-prompt detection. Contributed to implementation of feedback-driven conversational features, enabling the system to learn from user reactions to chatbot responses. Participated in a hackathon setting focused on improving prompt robustness through evaluation. • Implemented chatbot features such as like/dislike feedback and text-to-audio • Designed chatbot guardrails to constrain unsafe or invalid outputs • Created adversarial test prompts to improve invalid-prompt flagging accuracy • Supported model behavior improvement through iterative testing in a Graduate-Intern Hackathon

2023 - 2024

Education

Work History