Manasi Patwardhan

Product Engineer

Pune, Maharashtra, India24 yrs 7 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

Expert in Generative AI and Natural Language Processing.
Proven leadership in AI research and development.
Strong background in multi-modal reasoning and neural program synthesis.

Stackforce AI infers this person is a leading expert in AI and machine learning with a focus on research and development.

Contact

Skills

Core Skills

Generative AiNatural Language ProcessingMachine LearningSoftware EngineeringArtificial Intelligence

Other Skills

Research MentoringResearchCommunicationOriginal ResearchTransformersJavaCOBOLInductive Logic ProgrammingDeep LearningData MiningProgrammingData AnalysisData WarehousingDistributed SystemsAlgorithms

About

Senior Scientist at TCS Research. Leveraging my skills and creativity to seek possible opportunities for contributing as an AI and machine learning expert in the field of Generative AI, Large Language Models, Natural Language Processing and multi-modal data including code. Strong research, technical and analytical background Self-motivated, enterprising, fast learner, team oriented Proven leadership capabilities Excellent verbal and written communication skills Specialties: Expert in Natural Language Inference and Reasoning, multi-modal reasoning, neural program synthesis, AI4Code

Experience

24 yrs 7 mos

Total Experience

4 yrs 1 mo

Average Tenure

9 yrs 4 mos

Current Experience

Tata consultancy services

2 roles

Principal Scientist

Promoted

May 2025 – Present · 1 yr

Development of Acceleron - Gen AI powered Accelerated Research Platform
Semi-automating research stages such as ideation, exploration, experimentation, codification, communication and critique
Domain and Compositional Adaptation for NL-toSQL Semantic Parsing

Research MentoringGenerative AINatural Language Processing

Senior Scientist

Jan 2017 – May 2025 · 8 yrs 4 mos

Core Projects:
Java to COBOL code translation
Business Rule Extraction from COBOL Codes
Neuro-symbolic approach for Query Learning from few examples using Inductive Logic Programming for images and videos for domain dependent tasks
Advertisement Understanding for ad serving, creative compliance and retrieval: keyword / NLP based ad-retrieval setting (Can be generalized for Image retrieval)
Interpretability and reasoning using attention map analysis and probing
Fashion retrieval using latent style vector for outfit recommendation
Multi-lingual intent classifier using language model such as BERT
Modeled and designed algorithms for Samuhaa which is a machine with human-in-loop platform
AutoML: Neural Architecture Search for searching optimized Architectures suitable for a given dataset using reinforcement learning
Page-rank and LSTM based ranking for free-flowing textual annotations having inter-dependencies such as post-edit translations
Modeling crowd bias for successive tasks
Collaborations:
Material property prediction using neural network based techniques, Representation learning for atoms and compounds
Defining food digital signature for determining food freshness across food items using transfer learning and meta-learning techniques
Collection of regulation disambiguation data from crowd replacing experts. Expectation Maximization based crowd consensus, modeling parameters denoting worker behavior and skill, statement / phrase ambiguity, treating inter-annotator disagreement as an indication of ambiguity
Collection of data for research publication sentence classification based on intent for IPR chatbot, crowd-teaching set-up for teaching the crowd IP concepts

CommunicationOriginal ResearchTransformersNatural Language ProcessingMachine Learning

Vishwakarma institute of technology, pune

2 roles

Dean Research and Development and Industry Relations

Sep 2014 – Dec 2016 · 2 yrs 3 mos

Original Research

Professor

Jul 2011 – Dec 2016 · 5 yrs 5 mos

Member of the Academic Board as director’s nominee, Chairman, Board of Studies of Computer Engineering Program
Subject teacher for UG, PG and Ph.D. programs for Deep Learning, Machine Learning, Data Analytics, Business Intelligence, Data Warehousing and Mining, Object Oriented Modeling and Technology, Distributed Systems, Advanced Data Management Systems, Advanced Operating systems and Statistics & Research Methodology
Recognized Ph.D. Supervisor at SPPU and Nagpur University
Instrumental in working on funded research projects as a PI in the domain of Image and Video Processing, Social Media Mining, and NLP. Funded by DST, RGSTC, AICTE, SPPU, etc.
Ph.D. Student Projects:
◦ Non-normal Behavior prediction through online social network analysis
◦ Real-time Facial expression identification using video data
◦ Developing personalized digital signature using multi-task representation learning
◦ A formal cognitive emotion model for identification of music invoking soothing effect as a probing tool for meditation
◦ Video/ image mining for evaluation of aesthetics
Masters Student Projects:
◦ Context-based Citation Index of Research Articles using Graph Traversing and Text Mining Techniques
◦ Enhancing Sports Analytics using Time Series Summarization
◦ Analyzing the effect of twitter sentiments, commodity prices and analyst recommendations on Indian stock market
◦ Medical Side-Effect Analysis using Tweeter Feed
◦ Identification of Key Opinion Leaders in Healthcare using weighted Social Network Analysis
◦ Applying Swarm Intelligence approach for dynamic data mining
◦ Designing a framework for Automation of Educational Data Storage, Accreditation and Prediction (AEDSAP)
◦ Articulated Entity Relationship Diagrams (AER) for Complete Automation of Database Normalization
◦ ‘Smart Gram’: A presidential project for patient appointment scheduling system

Vishwakarma institute of technology, pune 411037

Associate Professor

Mar 2009 – Jun 2011 · 2 yrs 3 mos · Pune Area, India

Member of the Board of Studies of Computer Engineering Department as a subject expert.
Subject teacher for UG for Object Oriented Modeling and Technology, Distributed Systems courses and for PG for Advanced Data Management Systems, Advance Operating systems and Research Methodology courses
Guided UG and PG students for their thesis and seminars pertaining to various topics based on distinct computer science domains such as Data Models, Legacy Databases, Database Performance, Image Retrieval and Storage, Semantic Web Engineering, Web mining, image processing, etc.
Acting as an academic convener for collaborations with international universities such as Penn State University, USA, university of Alaska, Fairbanks, USA and colleges such as Marist College, USA.

Knowledge based systems inc

Researcher and Developer

Jul 2007 – Oct 2008 · 1 yr 3 mos · College Station, TX, USA

Researching and developing IDEF (Integration DEFinition Methods), BPEL (Business Process Execution Language) and UML (Unified Modeling Language) Models for Software Verification and Validation purposes
Static (Meta-Model compatibility) and dynamic (trace validation) checking of IDEF3 models including Process Flow Diagrams (PFDs) and Object State Transition Network (OSTN) using windows forms and C# code generation
Development of web-based applications in ASP .NET and C# including form based authentication, data validation and repair (server and client side controls), report generation, dashboard trend and metrics components (.NET Charting), SQL database access using data access layer components such as ADO .NET
SQL server database access using Structured Query Language (SQL)

University of tulsa

Graduate Research Scholar

Aug 2001 – May 2007 · 5 yrs 9 mos

Establishing a framework for assessment of distributed, dynamic and integrated System of Systems (SoS) from security perspective to facilitate certification Categorizing and distributing the security policy information specific to System of Systems across distinct policy types and descriptors
Modeling security issues and security certification concepts expressively and uniformly for software components using extensible, comparable and reusable modeling technology (UML Profile)
Applying the framework to examine scenarios of execution to determine policy mismatches and certification criteria violations using tool support
Assessing any negative impact on confidentiality, integrity and availability of the system, given a degree of violation
Research in software integration, middleware technology, grid security
Research in modeling technologies such as set theoretic notations, IO automata, pi-calculus, unity language, Unified Modeling Language, fault trees, etc
Responsible for software development in the field of integration of collaborative project management tools and developing a SQL database and its front end in MS Access for Center for Aircraft Support Infrastructure (CASI) Tinker Project