---
title: "Sushant Gundla - Technical Lead - AI ML and Innovations in Bengaluru | Tal"
description: "Meet Sushant Gundla, 20K Followers @LinkedIn ⭐ | Building Agentic Platforms, Chatbots, LLM Gateways & RAG Systems based in Bengaluru. AI stack includes claude code, hugging face, and langchain. Explore public work links including github, medium, and sushantgundla.com."
image: "https://tal.club/meet/sushant-gundla/share-image"
---

# Sushant Gundla on Tal

## Headline

20K Followers @LinkedIn ⭐ \| Building Agentic Platforms, Chatbots, LLM Gateways & RAG Systems

## About

I build AI systems that save teams thousands of hours — from Agentic RAG frameworks handling millions of records, to production LLM pipelines deployed across enterprise products at scale\. Senior AI Engineer at ConnectWise with 6\+ years of hands\-on experience in Generative AI, Agentic AI, NLP, and Computer Vision\. I specialize in turning complex AI research into practical, high\-impact systems that drive measurable business results — 60% reduction in ticket resolution time, 50% drop in support costs, and RAG apps serving every internal team\. What I work on: → Agentic AI frameworks \(multi\-agent orchestration, tool use, memory, planning\) → RAG & LLM pipelines \(OpenAI, LangChain, custom vector DBs, prompt engineering\) → NLP & Computer Vision \(transformers, fine\-tuning, domain\-specific LLMs\) → ML infrastructure \(FastAPI, AWS, Kubernetes, Docker, MLOps\) I also share weekly insights on Agentic AI, LLMs, and the open\-source AI ecosystem with 19K\+ followers on LinkedIn\. Open to senior AI/ML roles and collaboration on Agentic AI projects\. Feel free to message me here or connect directly\.

## Location

Bengaluru

## AI stack

claude code

hugging face

langchain

mcp

Agno

LiteLLM

## Work experience

Technical Lead \- AI ML and Innovations at PDI Technologies, Sep 2025 \- Present · 11 mo, Bengaluru

Senior Machine Learning Engineer II at ConnectWise, Feb 2024 \- Sep 2025 · 1 yr 8 mos, Bengaluru, Karnataka, India, 1\. Agentic Ticket Resolution Recommender Developed an Agents\-powered Ticket Resolution Recommender using the Agno framework, leveraging historical tickets \(millions of records\), device databases, IT documentation, RPA bots, and custom scripts\. The system reduced average resolution time by 60% and cut support team costs by 40%\. Built and deployed using FastAPI, OpenAI and AWS Kubernetes\. 2\. Agentic AI Framework Designed and implemented an Agentic AI framework for building and deploying config\-driven AI agents, accessible via REST API with support for: a\. Conversational b\. User\-specific memory c\. Multiple knowledge bases d\. Prebuilt agents for handling product\-specific interactions and automating common tasks e\. Integration with external tools such as web search, internal data access, and other enterprise systems 3\. In\-house RAG Framework Architected and developed an in\-house RAG framework for building low\-code, config\-driven RAG applications, enabling rapid deployment of RAG apps via REST API\. The framework: a\. Supports multiple LLM providers and vector DBs b\. Enables scalable ingestion of data like text, PDFs, CSV etc\. c\. Supports data filters, custom transformations, llm response formatting and much more d\. Prebuilt prompts for common use cases This framework is being actively used by developers and product managers of all internal teams 4\. RAG and Prompt Engineering Applications Built and scaled multiple RAG and Prompt Engineering applications using OpenAI models, serving use cases like: a\. RPA bot creation b\. Script Generation c\. Workflow Planning System d\. IT documentation generation e\. Documentation suggestions f\. RPA bot recommendation 5\. Internal Chatbot for Employee and Client Queries Developed an Internal Chatbot for automating employee and client queries, integrating knowledge sources such as HR policies, University resources, Virtual Community, Network/Cybersecurity, more reducing client support resolution time by 50%\.

Data Scientist III at Draup, Jan 2023 \- Feb 2024 · 1 yr 2 mos, Bengaluru, Karnataka, India, •Domain Specific Language Model \- Initiated pre\-training of in\-house Roberta\-based language model using 60M job descriptions, achieving performance gain over 5% on downstream tasks\. \- Implemented various techniques for efficient GPU training to reduce costs\. \- Conducted research on domain differences and optimized training parameters, including sequence lengths, optimizers, learning rates, schedulers, and batch sizes\. •Curie \- LLM RAG based Querying System \- Led the development of Curie, leveraging OpenAI ChatGPT, Langchain, and various prompting methods like CoT to analyze job descriptions\. \- Successfully scaled the system to respond to queries using 300M job descriptions in under 10 seconds\. \- Managed the project from concept to full\-scale production pipeline implementation\. \- Conducted multiple sessions on concepts of LLM Pretraining, Instruction Tuning and Parameter\-Efficient Finetuning of LLMs\.

Data Scientist II at Draup, Sep 2021 \- Dec 2022 · 1 yr 4 mos, Bengaluru, Karnataka, India, \- Developed Talent Report Recommendation system using collaborative and content\-based hybrid technique \- Designed advanced Skill Clustering algorithm with 93% top\-5 accuracy using Sentence Transformers based semantic similarity\. \- News Signal Classification •Diagnosed and resolved legacy issues, improving Data quality by architecturing new Data Workflow •Conducted 50\+ experiments for model fine\-tuning during continuous EDA\. •Monitored 8 production projects with adaptability to meet tight deadlines \- Demonstrated proactive leadership by driving projects into production and mentoring new team members

Senior AI/ML Devloper at Accolite Digital, Oct 2020 \- Sep 2021 · 1 yr, Bangalore Urban, Karnataka, India, \- Information Retrieval from Identity •Designed a Single Shot Object Detection\-based ID card localizer and Heatmap Regression\-based ID card KeyPoint Detector\. •Engineered SOTA Deep Learning models for ID card classification with 95% F1\-score •Implemented GPU\-based Image Processing pipeline for OCR optimization\. •Enabled real\-time Information Retrieval from 40\+ ID card types across 13 countries \- Back\-end and ML\-ops •Architected end\-to\-end system architecture and pipelines for ML micro\-services using Django and DL models version control •Utilized Docker and AKS for GPU Node deployment

AI/ML Engineer at Accolite Digital, Jul 2019 \- Oct 2020 · 1 yr 4 mos, Bangalore, •Developed CNN\-LSTM based model achieving 92% F1\-score for detecting Criticality in Radiology Report •Created LSTM based context\-based Profanity Detector •Implemented Mixed\-GANs for 15th\-century Avatar generation •Crafted Real\-time Face Matching with ID photos using Siamese model •Engineered real\-time Person Alive Detector using Facial Expressions and Pose detectors with sub\-100ms latency on mobile devices using TensorFlow\-JS

Associate Researcher at COEP \(College of Engineering, Pune\), Aug 2018 \- Jul 2019 · 1 yr, Pune/Pimpri\-Chinchwad Area

## Education

Master of Technology \- MTech, Computer Engineering \(Information Security\), COEP Technological University

Bachelor of Technology, Computer Engineering, KJ Somaiya College of Engineering, Vidyavihar

## Public links

- github: https://github\.com/sushantgundla
- medium: https://medium\.com/@sushantgundla
- sushantgundla\.com: https://sushantgundla\.com/

[View the original profile](https://tal.club/meet/sushant-gundla)

```json
{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "Person",
      "@id": "https://tal.club/meet/sushant-gundla#person",
      "name": "Sushant Gundla",
      "url": "https://tal.club/meet/sushant-gundla",
      "image": "https://tal.club/meet/sushant-gundla/profile-photo.jpg",
      "description": "Meet Sushant Gundla, 20K Followers @LinkedIn ⭐ | Building Agentic Platforms, Chatbots, LLM Gateways & RAG Systems based in Bengaluru. AI stack includes claude code, hugging face, and langchain. Explore public work links including github, medium, and sushantgundla.com.",
      "jobTitle": "Technical Lead - AI ML and Innovations",
      "worksFor": {
        "@type": "Organization",
        "name": "PDI Technologies"
      },
      "hasOccupation": {
        "@type": "Occupation",
        "name": "Technical Lead - AI ML and Innovations"
      },
      "homeLocation": {
        "@type": "Place",
        "name": "Bengaluru"
      },
      "alumniOf": [
        {
          "@type": "EducationalOrganization",
          "name": "COEP Technological University"
        },
        {
          "@type": "EducationalOrganization",
          "name": "KJ Somaiya College of Engineering, Vidyavihar"
        }
      ],
      "knowsAbout": [
        "claude code",
        "hugging face",
        "langchain",
        "mcp",
        "Agno",
        "LiteLLM"
      ],
      "sameAs": [
        "https://github.com/sushantgundla",
        "https://medium.com/@sushantgundla",
        "https://sushantgundla.com/"
      ],
      "mainEntityOfPage": {
        "@id": "https://tal.club/meet/sushant-gundla#webpage"
      }
    },
    {
      "@type": "ProfilePage",
      "@id": "https://tal.club/meet/sushant-gundla#webpage",
      "url": "https://tal.club/meet/sushant-gundla",
      "name": "Sushant Gundla on Tal",
      "description": "Meet Sushant Gundla, 20K Followers @LinkedIn ⭐ | Building Agentic Platforms, Chatbots, LLM Gateways & RAG Systems based in Bengaluru. AI stack includes claude code, hugging face, and langchain. Explore public work links including github, medium, and sushantgundla.com.",
      "dateModified": "2026-09-02T04:08:08.581Z",
      "isPartOf": {
        "@id": "https://tal.club/#website"
      },
      "mainEntity": {
        "@id": "https://tal.club/meet/sushant-gundla#person"
      },
      "primaryImageOfPage": {
        "@type": "ImageObject",
        "url": "https://tal.club/meet/sushant-gundla/profile-photo.jpg",
        "contentUrl": "https://tal.club/meet/sushant-gundla/profile-photo.jpg",
        "caption": "Sushant Gundla profile photo on Tal"
      },
      "inLanguage": "en-IN"
    }
  ]
}
```