Founded a consultancy specializing in AI and ML Engineering, providing strategic guidance as well as hands-on implementation to help companies build robust AI/ML solutions (with special focus on LLMs), including establishing the necessary data and infrastructure foundations.Partnering with diverse companies across industries, I gain exposure to cutting-edge technologies and evolving challenges. While I am fully enjoying the exposure and rapid learnings, I am open to exploring full-time opportunities with the right fit where I can continue to grow and contribute for the long-term.* Lead the development and implementation of an AI-powered medical calculator recommender engine, leveraging Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) systems to enhance clinical decision support for physicians.* Design and implement data preprocessing pipelines, vector database integrations, and evaluation frameworks to optimize system performance, delivering actionable medical calculator suggestions.* Collaborate with stakeholders/users on strategic planning, UX research, and technical documentation to align AI initiatives with company goals, improving both user experience and backend system efficiency.* Engineer and deploy AI-powered solutions for lease abstraction and invoice processing, integrating AWS tools and OpenAI to streamline document handling and enhance operational efficiency in property management.* Develop end-to-end API backend and VectorDB-based document retrieval systems, improving document parsing, database integration, and LLM-driven question-answering capabilities for property management applications.* Design and Implement ETL pipelines for client data to support generalized document extractions, lease abstraction, and accounts payable workflows, improving data standardization and system scalability.* Advise on Infrastructure-as-Code solutions (Terraform, CDK) for managing AWS assets, permissions/security, and deployments.

>> MLOps / ML Tooling / ML Infra / Product ML << • Integrate LLM solutions into product. Guide solution development to further automate legal workflows (reduce 6-figure onboarding costs to zero for clients). Rapidly implement eval and prompt management; design future architecture on common LLM services/libraries/tools. • Designed, developed, and launched a Service Mediator API gateway in prod for abstracting ML services, accelerating integration across four product areas by delivering predictions from 15 ML services via unified schema. Significantly boosted app engineering adoption, involving close collab with stakeholder teams in API design and architecture refinement • Assume pivotal responsibility for all PROD oversight of ML capabilities. Lead development for prediction delivery mechanisms and protocols to directly integrate with the user experience. Mentor and collab with Eng teams to handle prod workloads and achieve operational success • Drive service reliability and observability efforts: employ external tools (Datadog) and create self-hosted services (MLflow, streamlit) to enhance visibility of models. Created a portable load-test tool to allow stress testing our services in our environments. Implement monitors and alerts to ensure high operational standards and technical excellence • Automate ML Lifecycle phases to reduce time-to-production for model devs. Adopt IaC practices using Terraform/CDK for cloud management. Manage releases via CI/CD automation, containerization for service deployments (AWS, ECS, K8S) • Lead the successful deployment and manage hosting of on-prem Datasaur app to enhance data collection efforts, complete with secure auth and annotator management • Streamline team processes: Introduced MLOps Help Desk, office hours, and on-call rotations for stakeholder teams to ensure timely responses while fostering a more efficient & focused team environment to handle operational efforts, ad-hoc requests, and cross-team relationships

• Assumed leadership to expand Ontra's ML team, successfully growing the team from 4 to 10 teammates in 2022. Undertook managerial duties to guide and develop the team, while maintaining a high level of individual contributor work. Enabled the ML team to concentrate on product impact and advancing prediction coverage for NDA that accounts for 90% of the company's revenue • Initiated and advocated for integration of embedded ML resources on product teams to expand impact on more business areas, directly addressing core company challenges with ML solutions • Championed cross-team collaboration and fostered robust partnerships with product and engineering teams, enhancing stakeholder management, and upwards value communication • Implemented project management methodologies on AI initiatives, significantly improving team accountability and velocity. Established roadmap and planning processes to drive alignment • Accelerated ML team’s output and impact dramatically, evidenced by tripling the number of models in production from 4 to 12, supporting business objectives for reducing manual lawyer labor • Transitioned back to a Staff MLE role full-time, after ensuring leadership continuity by hiring a replacement Director. Concentrate on professional passions and deeply engage in solving technical challenges to develop and scale ML models, tools, infrastructure, and systems

Data Modernization @ Ontra.aihttps://www.builtinsf.com/2021/06/24/sf-data-modernization-success-incloudcounsel>> All things DataOps & MLOps! << • Operationalized ML efforts by developing a scalable prediction delivery system composed of clause detection and question-answer models to cover of all important aspects of NDA contracts • Spearheaded improvements to MLOps lifecycle including: model microservice ALB deployments, ci/cd testing & release, model artifact & dataset management, feature engineering, annotation ops • Created foundational data infrastructure in support of company-wide data analytics initiatives Cover all aspects including administering data warehouse (Snowflake), Airflow, ETL Pipelines • Drove Data Platform products: created microservices for data enrichment, nlp, and document processing capabilities to unlock pdf/docx functionality, entity extraction, and search • Orchestrated AWS resources via infra-as-code CDK (Cloudformation) and Gitlab ci/cd tools • Cultivated and led a team of 4 data engineers, lay foundations for scaling R&D hiring processes • Collaborated and managed complex cross departmental projects. Educated technical teams on best practices and instill agile processes at the rapidly growing company from 50 to 200 folks • Hold patent for an innovative system for constructing digital contract and applying ML modelsPython, Airflow, Snowflake, Postgres, Docker, AWS Ecosystem (CDK/CloudFormation, S3, ECS, Fargate, ECR, EC2, IAM, VPC, ALB, MWAA), microservices, RBAC, machine learning predictions, devops, github actions, gitlab ci/cd

* Mentoring and project planning, set strategic direction, and provide technical guidance for team* Recommendation and ranking systems, prototype machine learning solutions and deploy to production* Spearheaded innovative projects on better content understanding and graph structures

* Specialize in scaling ML prototypes to production systems* Designing and scaling content recommendation systems* Newsfeed ranking algorithms, Spark ETL pipelines* NLP, Topic Modeling, Classification

Summary:* Optimize for engagement metrics on user network and network activity* Modeling and statistical analysis.Details:• Spearhead topic modeling and content recommendation system for articles based on text content and historical user behavior• Specialize in scaling and designing machine learning systems; built ranking algorithm and ancillary infrastructure in Spark from ground up to serve content for users’ Newsfeed experience, revamping old default homepage experience with more engaging and personalized content • Architect data systems; created AB testing frameworks for network digest email and Newsfeed user layouts for split testing • Orchestrated efforts as technical lead for massive infrastructure enhancements to expand diversity of content types and boost speed performance. Interfaced with various stakeholders through effective communication and specifications to deploy features under tight deadlines • Deployed ranking system and algorithm for network digest email. Efforts included ETL of raw click data, feature engineering and model iteration; set up A/B testing framework to gather randomized signal and test various algorithms that combined business logic and models for soft rollouts that balanced delivery quantity with quality • Collaborated in initiatives for user growth and first-time user experience campaigns; performed ad-hoc analysis, data ETL and preparation, email targeting to increase registration rates for new users • Overhauled survey system for physician skills endorsements; created robust pipeline that gathered signals for physician skills from various data sources, including components that integrated human feedback to generate personalized survey nominations for survey takers • Introduced rigorous technical team standards and foundations to support fast team growth

• Designed data acquisition and ETL pipelines for sustainability, certification, and product data • Deployed internal Flask microservices on AWS using docker for automated product data ingestion, cleansing, processing, and storage; invented internal toolsets for convenient data extraction • Transformed and modeled unstructured data into various data stores (Neo4j, Postgres, MongoDB) • Researched and assessed viable datasets; worked with analysts on building an annotation pipeline • Performed exploratory analysis using clustering, word2ve, t-SNE visualization to assess similarities between products, product categories, certifications, and manufacturers • Generate insightful visualizations and reports to help stakeholders comprehend underlying trends

• Investigated relationships among 70,000 portfolios and 450,000 securities in the Bloomberg database for portfolio holdings recommendation and outlier detection on trading strategies • Implemented efficient similarity calculation of portfolio holdings with novel algorithm using binary compression and decompression for securities representation • Data Science, Machine Learning Algorithms, Classifiers • Business Intelligence, Workflow efficiency enhancements, automation • Adapted classification algorithm on data workflows, achieving automatic rejection rate of 45% • Designed business rule engine and customizable platform to generate data-driven news stories • Workflow efficiency enhancements, automation • Extraction, Natural Language Processing (NLP) • Natural Language Generation (NLG)

• Delivered marketing-moving headlines in subseconds using proprietary software; covered 200 tickers • Decomposed news into structured events with 20+ rulesets using entity extraction software (Inxight) • Earnings extraction and headline automation • Natural Language Processing (NLP) • News Extraction for Structured Feeds (Algorithmic Traders) for targeted topics • Workflow application development, automation with VBA/Excel/Macros.

RD&E Finance at Skillman (Consumer Products Company) • Spearheaded headcount consolidation reporting process, developed a Headcount BVA model • Streamlined business budgeting process through Allocation Model allowing greater visibility • Merited Gold Encore Award, recognized for leadership initiatives and business acumen • Collaborated with financial analysts on monthly close, global consolidation, SOX compliance • Performed monthly variance and ad-hoc analyses

• Exemplified strong interpersonal skills to provide courteous customer service to patrons

Shanghai, China • Transformed company website by enhancing visuals and streamlining navigation by flattening site structure • Analyzed competitor marketing strategies; developed product proposal documents and presentations