- Developing fined-tuned, context injected, and human reinforcement learning LLM-based solutions to extract information from documents (drug pre-authorization guidelines, criteria and their logical rules, document layout, and logical entailment between criteria) to infer decision trees that anable accurate patient pre-approval for treatments, sequences of fined-tuned LLMs with vector and graph databases, azure ML/AI ecosystem- Developing conversational agents to support the IT Help Desk in personalization of solutions and reducing the volume and repetition of tasks for live agents to resolve issues; fine-tuning pre-trained models and using vector databases for sources of knowledge for solutions, azure ML/AI ecosystem- Building and leading a team of data scientists on NLP/LLM/AI projects- Designing AI infrastructure for the company

- Created and lead the AI division, grew research and data science teams (x7 in 2 years)- Created and executed the company AI vision to secure and position Holler as a key player in Contextual AI, Contextual Content Recommendation and Delivery for our mobile and cloud-based platforms serving personalized contextual content to 1B daily messages on major messaging/social platforms- Created a portfolio of state of the art technologies including machine learning (ML), deep learning (DL), reinforcement learning (RL), network science (NetSci), graphs, natural language understanding (NLU, NLP)- Supported all other business units: sales, marketing, studio, and customer satisfaction in studies, descriptive + predictive analytic, discovering audience insights to drive business results- Represented Holler in various conferences and panels to position the company as a key player in the industry, senior technologist to investors to obtain 65M Series B- Successfully transitioned 100\% of research projects to engineering/product on product roadmap, all deployed and increasing ad revs, customer satisfactions, and brand recognition- Developed AI solutions for real-time incremental conversation understanding with higher-level semantic representation at atomic and compositional levels (intent identification, story understanding, context)- Developed AI solutions for real-time incremental participant affective state understanding (emotions, sentiment, mood, etc.)- Developed AI solutions for real-time multi-modal contextual content recommendation system on mobile and cloud- Developed hypergraph solutions for inference

- Developed big data and analytics solutions on Spark ecosystem (ML, DL, NetSci, graph) for real-time bidding to complex machine learning algorithms to make milliseconds decisions on audience for serving creatives- Lead, improved and grew 3 teams (15): 2 data engineering teams and 1 data science team- Developed and executed the data vision to improve ad revenues, reduce overall computational cost, improve stability and reliability of the entire data warehouse to near zero downtime, through analysis and strategic redesign of key components and data pipelines- Developed efficient robust and reliable big data solutions for hourly batch processing of TBs of collected and third-party DMP data- Developed solutions for real-time processing of >45B daily bid requests, to decide whether to bid in ad auction, bid price, and pace throughout the campaigns- Optimized our growing large data warehouse (700PB) to produce reports for campaign progress and analysis, monitor data processing, complex aggregations and data loading and transformations on AWS (Kinesis, Redshift, Aerospike, Dynamo) and Spark- Developed algorithms for user/audience modeling to identify the right audience that are most likely to respond to the campaign for one on one real-time creative recommendations (machine learning, topics modeling, clustering, recommender)- Worked cross-functionally with product and business teams to get products and features analyzed, designed, implemented, and shipped- Collaborated with business to investigate new ideas to improve key company KPIs, conduct the research and present result

- Developed and executed the vision for advanced next generation analytic tools for eDiscovery and litigation to maintain leadership in the e-discovery market segment- Lead and grew on/off-shore development teams (35)- Part of Business Development to expand the company to horizontal market segments while improving the vertical niches- Developed algorithms and solutions for NLP, ML, and dynamic network analysis for very large litigation datasets (TB to PB)- Developed next gen solutions for predictive coding transforming the eDiscovery industry in using ML instead of a manual coding process- Developed graph solutions for representing litigation data and discover insights as well as missing key facts