Jihar Gifari

Own projects end to end: feature engineering in Python, validation against held-out periods with financial backtests, then deployment on GCP with Git, automated retraining and monitoring. Built a Bayesian revenue model in Google Meridian that measures each channel's contribution and forecasts revenue under alternative budget plans.

Founded and lead an AI consultancy and products company, running a cross-functional team of 6 while staying hands-on technically. Deliver bespoke AI solutions for clients, including GenAI integrations into existing analytics products, and build AI products such as multi-step LLM agents over live commercial data.

Led an AI-powered image scoring system for cottages.com, delivering a measured +4% booking uplift on the worst-performing 20% of listings while cutting production costs by 50%. Built an automated reconciliation system for Wayland Additive, improving data accuracy and operational efficiency.

Built a churn prediction system on Southeast Asia's largest e-commerce platform, used by finance and marketing to guide budget and contributing to around 15% improvement in customer retention. Partnered with BI analysts and commercial stakeholders on large-scale customer and transactional datasets in SQL, and deployed models into production with BI engineers.

Worked in regulated financial services (securities and investments), leading end-to-end delivery of a Mutual Funds feature, building Customer Analytics dashboards in Tableau and running sentiment analysis on app reviews to inform strategy.

Delivered data science courses (Python, SQL, statistics, ML, Tableau) to individuals and corporate clients; authored an officially published textbook (ISBN: 978-623-7178-43-9).

Grade: Distinction

Activities and societies: Built a fine-tuned GPT-3.5 Umrah assistant with a ChatGPT-like interface on Streamlit; 1st of 15 finalists from 15 countries, out of about 1,000 entrants (April 2024).

MSc in Data Science with Distinction. Dissertation built an LLM assistant combining fine-tuning, prompt engineering, and corrective RAG with real-time retrieval, winning 1st place in an international AI competition.

Grade: Cum Laude, GPA 3.60/4.00

Activities and societies: Completed early, in 3.5 years.

Bachelor of Economics with Cum Laude honors and GPA 3.60/4.00. Econometrics thesis applied event study analysis and hypothesis testing to sectoral stock-index returns, later published.

Completed a Data Science and Machine Learning certificate program.