Dhilip S profile picture

Hello, I'm

Dhilip S

AI/ML Developer | Python Developer | Artificial Intelligence | Machine Learning | Gen AI | Agentic AI

Get To Know More

About Me

๐€๐ˆ & ๐Œ๐‹ ๐„๐ง๐ ๐ข๐ง๐ž๐ž๐ซ | ๐€๐ ๐ž๐ง๐ญ๐ข๐œ ๐€๐ˆ | ๐ƒ๐จ๐œ๐ฎ๐ฆ๐ž๐ง๐ญ ๐ˆ๐ง๐ญ๐ž๐ฅ๐ฅ๐ข๐ ๐ž๐ง๐œ๐ž | ๐‘๐€๐† | ๐‹๐‹๐Œ ๐€๐ฎ๐ญ๐จ๐ฆ๐š๐ญ๐ข๐จ๐ง

AI & ML Engineer with 1+ year of hands-on industry experience designing and deploying AI-driven, agentic automation solutions across document-intensive and enterprise workflows. I specialize in building scalable, production-ready AI systems using LLMs, RAG architectures, OCR, and Multi-Agent Frameworks to deliver measurable business impact.

๐€๐ˆ-๐ƒ๐ซ๐ข๐ฏ๐ž๐ง ๐ƒ๐จ๐œ๐ฎ๐ฆ๐ž๐ง๐ญ ๐ˆ๐ง๐ญ๐ž๐ฅ๐ฅ๐ข๐ ๐ž๐ง๐œ๐ž:

Designed and deployed Python-based AI/ML solutions leveraging LLMs, OCR, NLP, and RAG pipelines to automate document understanding, data extraction, classification, and decision workflows.

๐€๐ ๐ž๐ง๐ญ๐ข๐œ & ๐Œ๐ฎ๐ฅ๐ญ๐ข-๐€๐ ๐ž๐ง๐ญ ๐’๐ฒ๐ฌ๐ญ๐ž๐ฆ๐ฌ:

Built LangGraph-powered multi-agent workflows for POโ€“Invoice comparison, mismatch detection, and dispute identification, enabling high-accuracy automated processing.

๐„๐ง๐-๐ญ๐จ-๐„๐ง๐ ๐ˆ๐ง๐ญ๐ž๐ฅ๐ฅ๐ข๐ ๐ž๐ง๐ญ ๐๐ข๐ฉ๐ž๐ฅ๐ข๐ง๐ž๐ฌ:

Implemented integrated pipelines combining embeddings, vector databases, OCR preprocessing, and LLM reasoning, improving document intelligence and operational efficiency.

๐Œ๐จ๐๐ž๐ฅ ๐ƒ๐ž๐ฏ๐ž๐ฅ๐จ๐ฉ๐ฆ๐ž๐ง๐ญ & ๐…๐ข๐ง๐ž-๐“๐ฎ๐ง๐ข๐ง๐ :

Developed and fine-tuned forecasting, classification, NLP, and computer vision models, achieving up to 15% accuracy improvement and 40% reduction in manual review effort.

๐‘๐€๐† & ๐€๐ ๐ž๐ง๐ญ๐ข๐œ ๐‘๐€๐† ๐€๐ซ๐œ๐ก๐ข๐ญ๐ž๐œ๐ญ๐ฎ๐ซ๐ž๐ฌ:

Engineered context-aware retrieval systems for intelligent query answering, improving response relevance and reliability.

๐„๐ง๐ญ๐ž๐ซ๐ฉ๐ซ๐ข๐ฌ๐ž ๐€๐ฎ๐ญ๐จ๐ฆ๐š๐ญ๐ข๐จ๐ง & ๐๐ฅ๐š๐ญ๐Ÿ๐จ๐ซ๐ฆ๐ฌ:

Delivered secure, role-based automation solutions using Power Pages, Power Automate, SharePoint, Azure Entra ID, and UiPath, including end-to-end permit management systems.

๐‘๐ž๐œ๐ž๐ง๐ญ ๐€๐œ๐ก๐ข๐ž๐ฏ๐ž๐ฆ๐ž๐ง๐ญ๐ฌ:
  • Built Agentic AI and RAG solutions within a ๐‚๐ž๐ง๐ญ๐ž๐ซ ๐จ๐Ÿ ๐„๐ฑ๐œ๐ž๐ฅ๐ฅ๐ž๐ง๐œ๐ž (๐‚๐จ๐„) environment
  • Achieved up to 90% OCR extraction accuracy across structured, unstructured, and handwritten documents
  • Delivered AI solutions adopted in real business workflows
  • Earned UiPath Certified Professional โ€“ Agentic Automation Associate (UiPath-AAAv1)

I am passionate about building real-world, scalable AI systems that combine LLMs, Agentic Reasoning, Automation, and enterprise security. I enjoy working at the intersection of AI engineering, document intelligence, and intelligent automation, and I am driven to contribute to impactful, future-ready solutions.

Explore My

Experience

Experience

Junior Associate
KGiSL
Sep 2025 - Present ()

Trainee
KGiSL
Oct 2024 - Aug 2025 ()

Developer Intern
Bejawar Innovative Technologies
Jun 2023 - Feb 2024 ()

Education

B.E. Electronics and Communication Engineering
KGiSL Institute of Technology

My Expertise

Skills

Programming

Python
HTML
CSS
Javascript
SQL

AI & ML

Artificial Intelligence
Machine Learning
Computer Vision
Large Language Models
Generative AI
Agentic AI

Technologies

Cloud Deployment (GCP)
Flask
Linux
RPA

Browse My Recent

Projects

Digital Forms

KGiSL

Form Development: Designed and implemented Permit, Extension, and Closure forms in Power Pages with role-based access for Users, Approvers, and Admins.
Process Automation: Integrated Power Automate flows to trigger actions for data retrieval, submission, and updates in SharePoint lists.
Data Management: Configured SharePoint lists as the backend for storing permit requests, approvals, extensions, and closure details.
User Experience: Enabled pre-populated fields and automated workflows to reduce manual input and ensure accuracy across user roles.
End-to-End Workflow: Established a complete lifecycle for permit requestsโ€”from creation to approval, extension, and closureโ€”ensuring transparency and traceability.

Finance Agent - Agentic RAG Application

KGiSL

Finance Agent: Built an AI-powered retrieval-augmented generation application using LangGraph, LM Studio, ChromaDB, and Streamlit for finance-related document querying.
Dynamic Vendor Table Selection: Designed logic to auto-select vendor-specific tables in ChromaDB based on the vendor name present in the user query.
Contextual Document Retrieval: Implemented intelligent filtering to search only within vendor-specific documents and return precise, relevant responses.
No-Content Handling: Integrated fallback logic to detect non-relevant queries and respond with "No content is found" for unmatched document contexts.
LLM Integration with LM Studio: Connected the application to locally hosted LLMs via LM Studio to ensure secure, efficient, and cost-effective query processing.
Streamlit-based Interactive UI: Developed a simple, intuitive interface allowing users to upload multiple vendor SOW files and interact seamlessly with the AI.

Aztrabot and Aztracore

Bejawar Innovative Technologies

AztrBot and AztraCore Development: Created a real-time AI chatbot using GPT-4 and integrated multiple APIs, resulting in a 25% reduction in response time.
Integrated LLMs and Multiple APIs: Improved query precision by 30% by merging LLMs such as Mistral and GPT-4 with various APIs.
LangChain for LLM-powered Development: Utilized LangChain to streamline LLM-powered application development, reducing development time and boosting performance.
Implemented Aztracore Vector Database: Deployed the Aztracore vector database to enhance query understanding and retrieval efficiency.
Designed Dynamic User Interface: Crafted an engaging and user-friendly interface using HTML, CSS, JavaScript, and Flask to improve user experience.

Augmented Question Answering System for Biomedical Literature Retrieval using Language Models and PubMed Integration

KGiSL Institute of Technology

Keyword Generation and Article Retrieval: Employed NLP models to create keywords for PubMed searches, extracting relevant biomedical articles. Leveraged LLMs to synthesize concise answers.
Data Storage and Deployment: Stored interaction data in Firebase and launched the system with Streamlit, providing a user-friendly interface.
User Engagement: Streamlit's interface facilitated intuitive interaction with the AI assistant for biomedical information retrieval.

Open AI Chatbot

KGiSL Institute of Technology

ChatGPT Integration: Directed conversation logic using ChatGPT via HTTP requests.
Uipath Automation: Automated JSON deserialization and data handling with UiPath.
Document Generation: Streamlined conversion of ChatGPT responses from Word to PDF.
Efficient Workflow: Facilitated easy PDF downloads, improving user interaction and document management.

Spelling Correction Bot

KGiSL Institute of Technology

ChatGPT Integration: Leveraged OpenAI's ChatGPT API to detect and rectify spelling errors in text inputs via HTTP requests.
UiPath Automation: Automated text processing tasks, such as converting multiline text to a single line and managing text input/output, using UiPath.
Excel Data Handling: Streamlined text data processing by automating spell-checking and updating Excel files.
End-to-End Solution: Enhanced spelling correction accuracy and efficiency through integrated AI and automation tools.

Recognition & Awards

Achievements

Winner in Uipath Citizen Developer SKILL-A-THON - 2022

2022

Created three bots integrating ChatGPT with RPA and Excel, demonstrating innovative automation solutions.

Special Achievers - Annual Day - 2024

2024

Recognized for outstanding contributions and achievements throughout the academic year.

Highest Coursera Certificate Received - ECE Dept - Annual Day - 2024

2024

Acknowledged for obtaining the maximum number of Coursera certifications in the ECE department.

Winner Top2 (Good) in Mixed Signal SoC design Marathon using eSim and SKY130 - 2022

2022

Created a PWM Increment Circuit using eSim and SKY130, placing among the top performers.

Winner Top4 (Good) in Mixed Signal Circuit Design and Simulation Marathon using eSim - 2022

2022

Recognized for excellence in mixed signal circuit design and simulation using eSim tools.

Get in Touch

Contact Me