● Spearheaded the development of custom transformer models for information extraction in the cybersecurity domain.● Leveraged a custom contrastive transformer model to provide an ontology-agnostic solution that links text mentions to an authoritative biomedical knowledge base.● Developed a GPT-3.5 based solution that dynamically generates titles and definitions for user-selected text.

Led the data science team with a hands-on approach, driving significant improvements in client Googlesearch rankings through strategic SEO optimizations.Received praises from high-profile clients for rapidly delivering innovative NLP solutions that enhanced business operations:● Innovated a GPT and sentence-transformer based prototype for improving site navigability with automated internal linking and smart anchor text generation.● Engineered a solution that categorizes millions of top webpages of a corporate client into their internal taxonomy and pinpointed taxonomy gaps using Google search data.Acknowledged by the CEO for outstanding leadership, creative problem-solving, and commitment toprojects, highlighted in a company-wide commendation.

Engineering lead in multilingual NLP information extraction. Principal investigator in DARPA funded program that performs information extraction, event causality and text clustering.Served in multiple parallel capacities: ● Principal investigator and R&D lead in projects.● Primary developer and mentor of Raytheon BBN’s deep learning NLP R&D system, which had become integral to nearly all of Raytheon BBN's NLP projects● Collaborated with external stake-holders in Government, private organizations, and universities.

Creator, architect, and primary developer of Raytheon BBN's flagship deep learning R&D machine learning (ML) system for NLP, which served as the foundation system in multiple multi-million dollar NLP programs. The system builds upon PyTorch and transformer deep learning models to enable a variety of NLP tasks, such as information extraction, event extraction, relation extraction, clustering, coreference, named entity recognition and more.Principal investigator (PI) and engineering lead in multiple multi-million NLP programs. These programs range from multilingual information extraction, clustering to discover salient concepts in corpus, relation extraction, to prediction of future cyber security incidents.Served as consultant and mentor. Selected and served as NLP/ML expert advisor and judge for Raytheon Innovation Challenge, on topics regarding "AI, ML, and expert systems for national security". Mentored and worked with developers to perform deep learning R&D work in NLP programs.

Led R&D in an IARPA project to identify novel concepts from text in 2015, which predated the release of PyTorch. Designed an in-house cosine-similarity loss and coded from scratch a Python solution thatperforms gradient descent on word embeddings vectors to optimize the similarity loss.A core developer in Raytheon BBN’s winning event extraction systems in TAC-KBP 2014 and 2015.

Postdoc in Professor Dan Roth's CCG group, publishing three papers in top tier NLP conferences (ACL, EMNLP, CoNLL) on relation extraction and event causality extraction.