Apr 2025 – Present · 1 yr 2 mos · Arlington Heights, Illinois, United States · On-site

Apr 2024 – Apr 2025 · 1 yr · Durham, North Carolina, United States · Hybrid

• Designed and implemented a chatbot leveraging GPT-4o, large language models (LLMs),
• retrieval-augmented generation (RAG), and LangChain for enhanced conversational
• intelligence.
• Integrated advanced AI models, including those accessed via Azure OpenAI GPT-4o, into
• existing software, and developed new AI-powered applications, enhancing functionality,
• and introducing innovative features.
• Created an advanced conversational assistant built on LLMs, improving user interaction
• and satisfaction through accurate and contextually aware responses.
• Integrated Generative AI techniques, including Retrieval-Augmented Generation (RAG)
• and Named Entity Recognition (NER), into chat systems, utilizing LangChain and Azure
• OpenAI GPT-4o, to provide relevant and accurate information, improving user
• experience.
• Azure AI Foundry was utilized to assess various LLM models, leading to the selection of
• OpenAI GPT-4o after a thorough prompt-based evaluation.
• My role as an architect involved selecting the best-fit models and suggesting top free
• solutions for embedding generation and reranking.
• Developed an Agentic AI system to perform calculations on values extracted from a table
• Extracted information from diverse document formats using Azure OpenAI GPT-4o and RAG,
• refined outputs using various parsers, and summarized key sentiments and topics from user
• feedback, leveraging Python and Pandas for data manipulation.
• Managed project tasks and code using JIRA, GitLab, and Git, and documented project details in
• Confluence, ensuring efficient team collaboration and project execution. Data was stored and
• retrieved using PostgreSQL Vector DB.
• Classified types of questions to improve interaction and response accuracy.
• To address the rate limit error, particularly affecting summary generation, the optimal context
• sliding window technique was collaboratively selected and successfully implemented.

Aug 2022 – Apr 2024 · 1 yr 8 mos · Tampa, Florida, United States

• Led an initiative to significantly reduce ticket response and resolution times within
• ConnectWise by employing Python and Spyder to analyse ticket data. This involved
• identifying outlier tickets exceeding statistically defined thresholds, enabling targeted
• interventions and process improvements.
• Developed robust algorithms to pinpoint tickets with unusually long response and
• resolution times, using statistical methods to define "whiskers" and identify tickets falling
• outside these acceptable ranges.
• The analysis provided actionable insights into bottlenecks and inefficiencies, leading to
• streamlined workflows and improved service delivery.
• Spearheaded a project to analyse customer reviews using advanced Natural Language
• Processing (NLP) techniques, leveraging Python, Spyder, and Hugging Face.
• Employed state-of-the-art Language Model Models (LLMs) such as BERT, RoBERTa, and
• ALBERT to accurately extract and interpret customer sentiment from review statements.
• Enhanced the precision of sentiment analysis by refining pre-trained LLMs with custom,
• domain-specific data, improving the model's ability to understand nuanced customer
• feedback.
• Designed and developed an interactive Problem Management Dashboard using Python,
• Spyder, Django, and React to provide real-time insights into ticket outliers.
• Implemented sophisticated algorithms to identify outlier tickets based on multiple
• parameters, including partner, site name, and resource name, enabling proactive
• problem management.
• Developed a Python-based system using Spyder to extract and segregate chat times based
• on team, improving operational efficiency and resource allocation. This project enabled
• better understanding of chat support workloads and team performance.

Jan 2022 – Aug 2022 · 7 mos · California, United States

• Advanced Sentiment and Entity Recognition: Developed a sentiment analysis system
• using TextBlob, Naive Bayes, and AWS SageMaker BlazingText.
• implemented custom Named Entity Recognition (NER) with spaCy 3, enabling nuanced
• understanding of user feedback and data. This project leveraged Python and cloud-based
• NLP services for scalable data insights.
• Metrics data was calculated and it was given to the frontend to publish
• Cloud-Based Data Analysis: Experience working with cloud-based tools such as AWS Sage
• maker and Treasure Data, showcasing the ability to leverage cloud resources for scalable
• data processing and analysis.

Apr 2021 – Dec 2021 · 8 mos

• Automatic Lab Report Extraction: Designed and implemented an NLP pipeline to extract
• critical information from lab reports, delivering structured results to business
• stakeholders. This involved using Python, Dataiku for data orchestration, and Flask for API
• development, alongside Bootstrap, HTML, and CSS for a responsive front-end. Project
• management included cross-functional coordination, ensuring on-time delivery.
• Customer Data Matching: Enhanced Data Integration: Developed and optimized an NLP-
• based system for matching master data with survey data within Dataiku, significantly
• improving performance by reducing execution time from 24 hours to 3 hours. This project
• showcased expertise in data integration and performance tuning, utilizing Python and
• Dataiku's visual programming interface.
• Marketing Engagement Analysis: Predictive Scoring and Customer Segmentation:
• Developed a propensity scoring model to calculate engagement scores for diverse
• markets within the Treasure Data cloud platform, and implemented an RFM (Recency,
• Frequency, Monetary) model to identify high-value customer segments, using Python and
• Treasure Data’s query language.
• Skilful implementation of Large Language Models and Named Entity Recognition to
• extract and categorize key information from unstructured text data, demonstrating a
• deep understanding of advanced NLP techniques.
• Dataiku Workflow Optimization: Proven ability to significantly optimize data workflows
• within Dataiku, demonstrating proficiency in data processing, model deployment, and
• performance tuning.

Jun 2017 – Apr 2021 · 3 yrs 10 mos

• Project: AIAPI | Oct 2020 – Apr 2021 (HCL Automation COE)
• Applied NLP techniques (TFIDF, BOW) to match feature file and wadl files automatically
• Automatically updated the status in MYSQL
• Applied NLP techniques to match feature files and web scraped paths
• Project: ZTT | Apr 2020 – Sep 2020 (HCL Automation COE)
• Applied machine learning algorithms to identify the website actions like click, hover, button press
• Applied NLP techniques to extract the required text and developed automated test script for testing tools.
• Project: Image Cropping | Dec 2019 – Mar 2020 (HCL Automation COE)
• Identified the difference between two images and ignored the selected region
• Project: iCAP | Jan 2019 – Nov 2019 (HCL Mechanical COE)
• Successful in PDF Image extraction using OCR-Pytesseract, HOG Features Extraction and Entity recognition using Spacy
• Worked closely to extract BOM Table and Title Block data from Mechanical Drawing using OCR and XGBoost
• Applied Natural Language Processing (NLP) techniques to process the raw text & extract text from Readable PDF’s using Spacy NLP Entity Ruler
• Project: IDP Processing | May 2018 to Dec 2018 (HCL Mechanical COE)
• Extensively driven Readable pdf data extraction and displayed in csv format
• Handled day to day issues and fine tuned the applications for enhanced performance
• Project: OTD Dashboard | Oct 2017 - Apr 2018 (HCL Mechanical COE)
• Key involvement in analysis and prepared Gant Chart view by grouping and filtering.
• Built various graphs using Python matplotlib
• Building reusable assets & solutions for future business problems with AI programming
• Worked with Python libraries like NumPy, Pandas, Matplotlib to extract data and manipulate data for exploratory analysis
• Project: SR750 Dashboard | Jun 2017 - Sep 2017 (HCL Mechanical COE)
• Efficiently extracted text from readable pdf’s using pdfplumber and converted to a structured format

Oct 2014 – Apr 2017 · 2 yrs 6 mos

• Active involvement in Project Management like billing, resource assignations and delimitations
• Played a key role in SWRD and SWDD reviews as per DO-178B guidelines for G7K, KC390 Aerospace Software programs
• MRJ - Client UTC Bangalore - Code Review and DLD Review.
• 4G Development - Client UTC Bangalore - C, Cygwin, windriver, DOORs, Synergy CM & CR, Citrix, Boot and BSP Development, MPC 7448 Microprocessor, DMA Concepts, Marvell DMA Controller and RTOS.
• 8051 Programming.

Jul 2011 – Aug 2014 · 3 yrs 1 mo

• Effectively developed BSP software for SPDA module
• Managed entire projects lifecycles including Design, Development, and Deployment, Testing and Implementation and support
• Headed team of 30 direct engineers and 4 Project Leads for prioritizing assigned tasks & delivering assignments to increase the abilities and productivity
• Accountable for preparation & implementation of HSIT, SSIT and LLT Test cases for WELS, DGCU Aerospace Software programs
• Aided Boeing 787 Dreamliner as a CCS lead, performed migrations for different block points at Dreamliner Lab at Rockford, IL
• Primarily involved in test case preparation & execution of LLT or HSIT and debugging to deliver the product within stipulated time
• Involved in test case preparation & execution of LLT for A400M MMU at client place i.e., at Cottbus Germany

Aug 2007 – Jun 2011 · 3 yrs 10 mos

• C, C++, Python, TClTk, LDRA, VectorCast, Boeing Maintenance Tools. DO-178B.

Sep 2005 – Jul 2007 · 1 yr 10 mos · Bangalore

• C, Win32