Sep 2024 – Aug 2025 · 11 mos · Bengaluru, Karnataka, India · Hybrid

• Worked as Software Engineer in Data Science Department of Organization
• 1) Legal Chatbot (MVP Product for Client): Developed a chatbot that connects users to lawyers based on query context and location using Python, GraphQL, Deepval, and Generative AI. Enabled intelligent matching and conversational flow.
• 2) Conversational AI Agent for Payment Collection: Built an AI agent that auto-calls customers for payment dues, understands responses, captures relevant details, and stores them in a database. Used Python, LangChain, LangGraph, Generative AI, and Ngrok for secure real-time communication.
• 3) Audio Analytics Platform: Designed a system to upload customer service calls, summarize conversations, and evaluate service quality using audio metrics. Employed Librosa and Generative AI to derive interaction insights and smoothness scores.
• 4)Dental Disease Detection System: Fine-tuned YOLOv8 using LoRA and QLoRA for efficient and memory-optimized classification of dental diseases from X-rays. Integrated RAG (Retrieval-Augmented Generation) to allow junior doctors to query about diseases from relevant docs and generate diagnosis transcripts.
• 5)Fraud Detection using LLM Fine-Tuning: Fine-tuned a Large Language Model using LoRA/QLoRA on transaction and support ticket data to detect fraud patterns. Integrated into the client’s backend to flag anomalies with contextual explanations.
• 6)Client Conversation Insight Tool (Internal): Developed a system to process continuous conversations between clients and PMs/analysts, extracting structured data and generating reports.
• 7)AI Interviewer (Internal): Built an end-to-end interviewer AI that evaluates candidate answers, assigns scores, and creates feedback reports autonomously.

Nov 2023 – Aug 2024 · 9 mos · Bengaluru, Karnataka, India · On-site

• Worked on Product InCbot an AI Interviewer
• 1) AI Interviewer Platform:
• Built InCBot, an AI-powered interviewer that generated DSA and behavioral questions, evaluated responses, and simulated interviewer interaction using LLMs and prompt engineering.
• Developed face verification API (YOLOv5 + OpenCV + MediaPipe) to ensure candidate identity validation.
• Used MediaPipe for real-time “looking away” detection, improving the accuracy of proctoring reports.
• Integrated DeepFace for facial emotion analysis and confidence scoring during interviews.
• Created audio-video merging pipeline using Wav2Lip and Whisper to generate AI-powered talking avatars.
• Designed an ETL pipeline for retraining communication and fluency scoring models using fine-tuned HuBERT.
• 2)AI-Powered Feedback Report System:
• Implemented a RAG (Retrieval-Augmented Generation) system to evaluate candidate responses against AI-generated expected answers using LangChain + LLaMA.
• Added AI-generated feedback comments for every candidate answer to provide contextual, semantic evaluation in reports.
• Integrated prompt-based compiler testing for automatic code evaluation and DSA feedback.
• Built an OCR + Regex parser to extract emails and phone numbers from candidate resumes in under 500ms.
• Developed REST APIs using Flask, containerized with Docker, and tested via Postman.
• 3)JD–CV Matching & Dynamic JD Generation :
• Built a Job Description to CV matching system using NLP and LLM embeddings to rank candidate profiles.
• Designed an LLM-powered dynamic JD generator that created job descriptions from minimal inputs (e.g., job role, years of experience, skills).
• Used techniques like prompt tuning and few-shot examples to improve generation accuracy and domain alignment.
• 4) Leadership & Coordination:
• Coordinated with QA and Product teams to gather requirements and translate them into AI features.
• Reported directly to the Product Manager and resolved issues raised by the Operations and Client Support teams.

Sep 2023 – Nov 2023 · 2 mos · Bengaluru, Karnataka, India · On-site

• 1)Worked on Computer Visions Models for Multiple face detection and object detection here I used pretrained models and Lightweight Model like MobileNet
• 2)Worked for multiple voice detection using speaker diarization
• 3)Use Pynnote for voice segmentation and speaker diarization
• 4)Worked on Avatar takative face geeneration for AI face by using models like Wav_2_LipGAN Model
• 5)Worked on Accuracy check of various Speech to text models
• 6)Worked on Cv and Resume parsing by using regular expression to extract information from Resume for that I used testract library of Python

May 2023 – Oct 2023 · 5 mos · Pune, Maharashtra, India · Remote

• TCS TAAP YEP program

Jan 2023 – Jul 2023 · 6 mos · Nagpur, Maharashtra, India

Jul 2022 – Aug 2022 · 1 mo · Online · Remote

• JAVA Coding Internship