Get in Touch

• Designed and led development of ALAI (Active Listening AI), a multi-component LLM platform processing financial advisor–client call transcripts using LangChain/LangGraph + AWS Bedrock (Claude). Built four production modules: General Summary (Zoom-style), FactFinder (comprehensive case notes with extended thinking), Discovery Letter (client follow-up emails), and Fact Extraction (93.4% field-level recall with Pydantic schema enforcement and ReAct self-correction).
• Designed ALAI Insights, a multi-agent research framework implementing scientific methodology (questions → hypotheses → investigations → insights → validations → conclusions) with LangGraph orchestration. Features universal voting for community consensus, entity citations for knowledge graphs, cross-run continuity, and progressive context management (200K → 1M token windows).
• Sole developer of a two-stage financial planning assistant: FT-Transformer for product propensity prediction and FiLM-conditioned transformer for page prediction. Implements plan “inpainting”—predicting optimal plan composition from client features and partial selections—trained on 1.2M+ historical plans with a weighted ensemble approach where product propensities inform page recommendations.
• Built multi-agent evaluation framework for fact extraction using synthetic SME evaluations; orchestrates specialized agents for field-level accuracy, schema validation, and quality metrics analysis across 14+ dimensions to optimize prompt engineering iterations.
• Evaluated transcription service alternatives to AWS Transcribe, contributing to migration recommendation that achieved ~60% cost reduction (~$300K annual savings).

• Collaborated on a clothing recommendation system using Doc2Vec embeddings within CatBoost, doubling the take rate for recommended items compared to non-recommended ones.
• Implemented algorithm reducing warehouse associates’ travel distance for order picking by 24%+.
• Built a model predicting order items that will receive fulfillment via returns within 24 hours.
• Created ‘Visual Sales’ dashboard returning a grid of images ordered by best sellers, using Shiny in R.

• Worked cross-functionally to extract, transform, and load (ETL) customer data into Penta Technologies ERP database.

• Constructed predictive models using supervised machine learning algorithms to predict patient mortality risk.

• Led a remodel program comprising 348 plumbing supply stores across the US and expedited the timeline by 75%+.

• Designed and built an autonomous RL-based portfolio management system that ingests multi-source financial data (FRED, SEC, news, market), processes it through 15 specialized feature engineering modules, and outputs risk-managed portfolio allocations using PPO with a Transformer policy network.
• Implemented cross-asset attention mechanism, market regime detection, and reward shaping balancing return incentives against drawdown and turnover penalties.
• Integrated as a tool within an LLM-orchestrated agent: the model surfaces insights and allocation recommendations, discusses rationale with a human-in-the-loop, then executes approved actions.

• Built a knowledge graph–based task management system where a single logged action recursively propagates through a multi-parent taxonomy spanning 6 category hierarchies and 1,600+ nodes.
• Engineered adaptive frequency system using golden ratio–based adjustments: tasks completed early tighten in frequency, tasks completed late loosen — the system evolves to match actual behavior without explicit configuration.
• Trained prediction model achieving ~75% Hit@1 accuracy for next likely action, served as a tool to an LLM agent that interprets predictions, recommends scheduling changes, and executes adjustments after human-in-the-loop approval.

Summa cum laude

Cum Laude · Swimming & Diving Team