Mukesh Reghu

AI Researcher

Chennai, Tamil Nadu, India5 yrs 11 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

Expert in multilingual NLP systems and LLMs.
Proven track record in AI-driven healthcare solutions.
Strong background in explainable AI methodologies.

Stackforce AI infers this person is a SaaS-focused AI specialist with expertise in NLP and deep learning.

Contact

Skills

Core Skills

Large Language Models (llm)Machine LearningNatural Language Processing (nlp)Deep Learning

Other Skills

Retrieval Augmented Generation (RAG)Multilingual TransformersConversational Sentiment Analysis (Turn, Call-level)Recurrent Neural Networks (RNN)Convolutional Neural Networks (CNN)Cluster AnalysisExplainable AIAWS Certified Solutions ArchitectCloud ComputingAI AgentsOpenAI Function callingReAct PromptingComputer ScienceResearchProgramming

About

Experience

5 yrs 11 mos

Total Experience

2 yrs 11 mos

Average Tenure

3 yrs 7 mos

Current Experience

Uniphore

2 roles

Sr. AI Scientist

Promoted

Aug 2024 – Present · 1 yr 10 mos

◦ Built and productionized a new multilingual intent detection system for a real-time agent assist product, replacing a legacy system in production.
◦ Benchmarked the intent detection system against the incumbent across accuracy, latency, concurrency, and cost using internal customer POC data and public datasets; presented results to product managers and stakeholders.
◦ Led production integration end-to-end for the intent detection system, supporting QA, release, and enabling sales and delivery teams with customer demos.
◦ Extended the intent detection system system to multiple languages and applied LLM-based techniques to accelerate training and test data creation.
◦ Designed and evaluated LLM-based approaches for predicting the disposition path/leaf for a given call from a tree of hierarchical dispositions, such as recursive prompting and candidate retrieval followed by structured reasoning, achieving 47% exact-match accuracy on a customer dataset.
◦ Extended a production RAG system to support Arabic by benchmarking embeddings, re-rankers, prompts and LLMs using language experts and LLM-as-a-judge. Improved the semantic retrieval in the system by over 10%.
◦ Built an LLM-based canonicalization system to normalize semantically equivalent texts into a single canonical form, enabling reliable grouping, reporting, and insights; handled cross-pod race conditions using Redis-based distributed locks.

Large Language Models (LLM)Retrieval Augmented Generation (RAG)Machine Learning

AI Scientist

Nov 2022 – Aug 2024 · 1 yr 9 mos

◦ Built and owned a multilingual sentiment analysis system (turn-level and call-level) for telephony conversations, supporting 8 Indian and international languages, from annotation strategy to production deployment.
◦ Designed a novel low-data, multilingual approach for call-level sentiment prediction, achieving 0.866 accuracy on production data in the field.
◦ Mentored team members in extending the turn and call-level sentiment prediction system to multiple Indian and international languages.
◦ Developed and maintained an internal python library for text classification problems.

Multilingual TransformersConversational Sentiment Analysis (Turn, Call-level)Natural Language Processing (NLP)Machine Learning

Buddi ai

4 roles

Senior Research Scientist I

Promoted

Jul 2022 – Nov 2022 · 4 mos

Recurrent Neural Networks (RNN)Convolutional Neural Networks (CNN)Deep LearningNatural Language Processing (NLP)

Research Scientist II

Jul 2021 – Jun 2022 · 11 mos

Worked on NLP + deep learning systems for healthcare revenue cycle automation (EHR understanding, automated medical coding, and explainability).
◦ Experimented with various deep learning model architectures and techniques for speeding up the training and improving the performance of the system of deep learning models that aid the process of automated medical coding (including multi-GPU training and simple experiments with distributed model training)
◦ Researched on and delivered business solutions for explaining deep learning models for Named Entity Recognition (NER) and text classification that utilized popular explanation methods such as LIME, saliency maps, attention plots and Concept Activation Vectors (CAVs) in addition to a proprietarily developed method
◦ Exported and deployed deep learning models using the ONNX framework

Research Associate

May 2020 – Jun 2021 · 1 yr 1 mo

Worked on various NLP based deep learning models for performing tasks related to automated medical coding on Electronic Health Records (EHRs)
◦ Worked on the problem of multi-label long text classification with many labels
◦ Packaged the entire pipeline of training, evaluating, inferring, explaining the predictions of and exporting the system of models for automated medical coding as a python package (and command line tool) for use by other researchers and novice users

Research Intern

May 2019 – Oct 2019 · 5 mos · Chennai, Tamil Nadu, India

Internship and Final Year Project
◦ Fine-tuned BERT model further pretrained on biomedical data for various tasks such as Sentence Similarity, Named Entity Recognition, Relation Extraction, Document multi-label classification and Natural Language Inference
◦ Graph Neural Networks for NLP. Final Year Project Short Report: https://drive.google.com/file/d/12dNZYI9Z1T1aE_vXlxovMjImA9YUM2Ox/view?usp=sharing · Final Year Project Detailed Report: https://drive.google.com/file/d/1ZPIrKNsHADSYuuQBooMvif-B4CPzV6Xy/view?usp=sharing

Cluster Analysis

Wipro technologies

Project Trainee

May 2018 – Jul 2018 · 2 mos · Bengaluru, Karnataka, India

Buddi ai

Intern

May 2017 – Jun 2017 · 1 mo · Chennai · On-site

Worked with various machine learning algorithms/techniques, esp., on data pre-processing, clustering, cluster validation and also their implementations in Python