May 2025 – Jul 2025 · 2 mos · United States · Remote

• AI started expending with the help of GPT-2, GPT-3 which was mindblowing but hallucinating, guessing and mimicking.
• Then came the next phase: Reinforcement Learning (that’s where i started my carrer at open AI from Scale AI)
• I was there with 1600 other contributers in the loop evaluating responses. You can think of it as voting on intelligence and many peoples were doing poor work which was causing LLms to not performing well at OpenAI (at that time around 1200 contributers were kicked out from the organization and leaders added the recheck loop), which gave me another oppportunity of becoming a squad team memeber with rechecking work and training AI I was looking at 32 members ensuring they don’t submit bed work and we can maintain the quality of training.
• And things got Drastically changed when we stopped telling AI on what’s right and started letting it experience what experts actually do. That shift started the O-series (O1 and O3 models) and the thinking and reasoning models that changed everything.
• On the organization side Meta had invested on scale AI which caused Open AI to give less work and in parallel they were making the Agent systems and reducing the human loops
• and they implemented systems to train LLMs inside virtual machines for coding, breaking, That’s the foundation of modern agent systems where models can plan, reason, and act.
• During that time I was Selected as an Oracle Elite Contributor (to recognise top contributors who consistently demonstrate excellence in their work, allowing me to leverage my expertise while enjoying enhanced benefits and opportunities within the Oracle ecosystem.) And i contributed to Reinforcement Learning from Human Feedback (RLHF) and AI self-improvement systems.
• As the next phase began, OpenAI sunset its work through Scale AI, moving toward fully autonomous agent systems causing me to sift at Meta and got some issues there and decided to start JusTakeMyCard

Dec 2024 – Jul 2025 · 7 mos · Remote

• Evaluating and improving the quality of AI interactions, providing actionable insights for AI training and development.

Oct 2024 – Jul 2025 · 9 mos · Remote

• Work on Training LLMs with Reinforcement Learning with human feedback.

Sep 2024 – Oct 2024 · 1 mo · Davis, California, United States · Remote

• Working as a Data Science and AI Developer to create a global Scientific Research Community

Apr 2024 – Jul 2024 · 3 mos · India · Remote

• 𝐂𝐫𝐞𝐚𝐭𝐞𝐝 𝐂𝐨𝐦𝐩𝐫𝐞𝐡𝐞𝐧𝐬𝐢𝐯𝐞 𝐂𝐡𝐚𝐭𝐛𝐨𝐭𝐬: Successfully developed four chatbots tailored for different roles: Admin, Dealer, Collector, and User. These chatbots enhance user interaction by providing real-time information and personalized responses.
• 𝗢𝗽𝘁𝗶𝗺𝗶𝘇𝗲𝗱 𝗔𝗣𝗜 𝗦𝗽𝗮𝗰𝗲 𝗨𝘁𝗶𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻:
• 𝙄𝙣𝙞𝙩𝙞𝙖𝙡 𝙎𝙩𝙖𝙩𝙚: Managed four different APIs, each consuming approximately 2.3 GB, totaling around 6.9 GB.
• 𝙊𝙥𝙩𝙞𝙢𝙞𝙯𝙖𝙩𝙞𝙤𝙣: Consolidated functionalities into a single API, reducing the total space to 365 MB.
• 𝙍𝙚𝙨𝙪𝙡𝙩: Achieved a significant reduction of ~6.5 GB, leading to cost savings and improved performance.
• 𝐂𝐥𝐨𝐮𝐝 𝐄𝐟𝐟𝐢𝐜𝐢𝐞𝐧𝐜𝐲:
• Migrated from using three AWS instances to a single API running on a free trial part, optimizing resource usage.
• Transitioned from open-source models to AWS Bedrock models, ensuring better integration and efficiency.
• 𝐄𝐧𝐡𝐚𝐧𝐜𝐞𝐝 𝐂𝐡𝐚𝐭𝐛𝐨𝐭 𝐅𝐮𝐧𝐜𝐭𝐢𝐨𝐧𝐚𝐥𝐢𝐭𝐢𝐞𝐬:
• 𝙁𝙤𝙧 𝘼𝙙𝙢𝙞𝙣𝙨, 𝘿𝙚𝙖𝙡𝙚𝙧𝙨, 𝙖𝙣𝙙 𝘾𝙤𝙡𝙡𝙚𝙘𝙩𝙤𝙧𝙨:
• Access to real-time car information from the website using natural language processing.
• Minimized hallucinations by ensuring precise data retrieval compared to conventional models.
• Personalized interaction based on individual user history, such as purchase history, bidding information, user visits, and transaction history.
• 𝙁𝙤𝙧 𝙐𝙨𝙚𝙧𝙨:
• Easy access to website and car-related information, enhancing user experience and engagement.
• 𝗖𝗼𝗺𝗽𝗿𝗲𝗵𝗲𝗻𝘀𝗶𝘃𝗲 𝗜𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 𝗔𝗰𝗰𝗲𝘀𝘀:
• Enabled chatbots to provide details about website policies, cookies, privacy policies, and more, ensuring users have complete transparency and access to relevant information.

Feb 2024 – Mar 2024 · 1 mo · India · Remote

• This is a nice AI&ML-powered system with personalized diet planning and interactive personal answering capabilities by taking small information. This chat-bot is created on the following Things:
• 𝟏. 𝐏𝐞𝐫𝐬𝐨𝐧𝐚𝐥𝐢𝐳𝐞𝐝 𝐝𝐢𝐞𝐭 𝐩𝐥𝐚𝐧: ML-based Model which takes very less information from user and recommend diet to the person.
• 𝟐. 𝐏𝐞𝐫𝐬𝐨𝐧𝐚𝐥 𝐀𝐧𝐬𝐰𝐞𝐫𝐢𝐧𝐠: It has capability to give personal answers also for the user query.

Nov 2023 – Mar 2024 · 4 mos · India · On-site

• Developing an Ai-driven interview optimization system dedicated to the interview process, focusing on reducing interview time by 30%. This system will try to revolutionize the interview process and increasing interviewe Process by 15% upon full deployment. This solution will enhance candidate selection efficiency.
• KEY FEATURES :
• 1. 𝐑𝐞𝐬𝐮𝐦𝐞-𝐁𝐚𝐬𝐞𝐝 𝐐𝐮𝐞𝐬𝐭𝐢𝐨𝐧𝐢𝐧𝐠: Intelligent questioning based on candidates resumes to understand their current knowledge.
• 2. 𝐒𝐨𝐟𝐭 𝐒𝐤𝐢𝐥𝐥𝐬 𝐓𝐞𝐬𝐭𝐢𝐧𝐠 𝐏𝐫𝐨𝐠𝐫𝐚𝐦: Designing soft skills testing program for real-world scenarios. This initiative will score Candidate ability to ajust on the company.
• 3. 𝗣𝗿𝗼𝗷𝗲𝗰𝘁𝘀-𝗕𝗮𝘀𝗲𝗱 𝗤𝘂𝗲𝘀𝘁𝗶𝗼𝗻𝘀 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻: Asking project-related questions to candidate and evaluation of their skills as presented in their resumes.

Aug 2023 – Sep 2023 · 1 mo · India · On-site

• During my internship, I successfully developed and deployed multiple chatbots using a variety of innovative methods and cutting-edge technologies. Key highlights of my experience include:
• 𝙍𝙪𝙡𝙚-𝙗𝙖𝙨𝙚𝙙 𝘾𝙝𝙖𝙩𝙗𝙤𝙩𝙨: Designed and implemented chatbots using Regular Expressions to efficiently handle predefined queries and responses.
• 𝙉𝙖𝙩𝙪𝙧𝙖𝙡 𝙇𝙖𝙣𝙜𝙪𝙖𝙜𝙚 𝙋𝙧𝙤𝙘𝙚𝙨𝙨𝙞𝙣𝙜 (𝙉𝙇𝙋): Utilized NLTK (Natural Language Toolkit) to develop chatbots with advanced language processing capabilities.
• 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝘁 𝗜𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 𝗥𝗲𝘁𝗿𝗶𝗲𝘃𝗮𝗹: Leveraged TF-IDF (Term Frequency-Inverse Document Frequency) for keyword-based query handling, enhancing the chatbots' ability to understand and respond to user inputs.
• 𝗔𝗱𝘃𝗮𝗻𝗰𝗲𝗱 𝗡𝗟𝗣 𝗧𝗲𝗰𝗵𝗻𝗶𝗾𝘂𝗲𝘀: Applied sophisticated techniques such as streaming and lemmatization to improve chatbot performance and accuracy.
• 𝐈𝐧𝐭𝐞𝐧𝐭 𝐑𝐞𝐜𝐨𝐠𝐧𝐢𝐭𝐢𝐨𝐧 𝐚𝐧𝐝 𝐔𝐭𝐭𝐞𝐫𝐚𝐧𝐜𝐞 𝐇𝐚𝐧𝐝𝐥𝐢𝐧𝐠: Developed chatbots capable of accurately recognizing user intents and handling diverse utterances to provide relevant responses.
• 𝐔𝐬𝐢𝐧𝐠 𝐕𝐨𝐢𝐜𝐞-𝐟𝐥𝐨𝐰 𝐟𝐨𝐫 𝐀𝐈 𝐈𝐧𝐭𝐞𝐠𝐫𝐚𝐭𝐢𝐨𝐧: Created a cutting-edge chatbot using Voiceflow with ChatGPT integration, delivering a seamless and engaging conversational experience.
• This experience allowed me to hone my skills in chatbot development and gain hands-on experience with various NLP and AI tools, significantly contributing to my professional growth.

Feb 2023 – Aug 2024 · 1 yr 6 mos · Bhopal, Madhya Pradesh, India · On-site

• I am currently enhancing my skills and knowledge through a comprehensive learning program at Cybrom, covering a wide range of advanced topics in technology and data science. This program includes:
• 𝐂𝐨𝐦𝐩𝐮𝐭𝐞𝐫 𝐕𝐢𝐬𝐢𝐨𝐧:Understanding and applying techniques to analyze and interpret visual data.\
• 𝗡𝗮𝘁𝘂𝗿𝗮𝗹 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗣𝗿𝗼𝗰𝗲𝘀𝘀𝗶𝗻𝗴 (𝗡𝗟𝗣): Learning methods to process and analyze large amounts of natural language data.
• 𝗗𝗲𝗲𝗽 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴: Exploring neural networks and their applications in solving complex problems.
• 𝐓𝐚𝐛𝐥𝐞𝐚𝐮 𝐚𝐧𝐝 𝐏𝐨𝐰𝐞𝐫 𝐁𝐈: Gaining proficiency in data visualization tools to create interactive and insightful dashboards.
• 𝐁𝐢𝐠 𝐃𝐚𝐭𝐚 𝐰𝐢𝐭𝐡 𝐒𝐩𝐚𝐫𝐤: Utilizing Apache Spark for big data processing and analytics.
• 𝐁𝐢𝐠 𝐃𝐚𝐭𝐚 𝐰𝐢𝐭𝐡 𝐇𝐚𝐝𝐨𝐨𝐩: Learning the Hadoop ecosystem for scalable and distributed data processing.
• 𝐌𝐚𝐜𝐡𝐢𝐧𝐞 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠: Diving deep into machine learning algorithms and their advanced applications.
• 𝐋𝐢𝐧𝐮𝐱: Understanding the Linux operating system and its applications in various fields.
• 𝐃𝐚𝐭𝐚 𝐀𝐧𝐚𝐥𝐲𝐬𝐢𝐬 𝐔𝐬𝐢𝐧𝐠 𝐏𝐲𝐭𝐡𝐨𝐧: Developing skills to analyze data and derive meaningful insights using Python.
• 𝐑𝐃𝐁𝐌𝐒 (𝐒𝐐𝐋): Learning relational database management systems and SQL for effective data management.
• 𝐀𝐝𝐯𝐚𝐧𝐜𝐞𝐝 𝐄𝐱𝐜𝐞𝐥: Enhancing proficiency in Excel for advanced data analysis and reporting.
• 𝐏𝐲𝐭𝐡𝐨𝐧 𝐟𝐨𝐫 𝐃𝐚𝐭𝐚 𝐒𝐜𝐢𝐞𝐧𝐜𝐞: Utilizing Python programming for data science applications.
• Through this structured learning path, I am equipping myself with the necessary skills to excel in the ever-evolving field of technology and data science.

Feb 2023 – Apr 2023 · 2 mos · India · Remote

• Automated Video Upload System: Created a magical auto video taker and uploader system, streamlining the process of transferring content from Instagram to YouTube.
• Increased Engagement: Enhanced viewer engagement by 12% through more consistent and timely content uploads.
• Development Approaches:
• Frontend Automation: Developed a frontend solution using pyautogui, requiring a desktop or other media to automate the video selection and upload process.
• YouTube API Integration: Implemented a backend solution using YouTube posting APIs to directly upload videos, bypassing the need for manual intervention.
• Playwright Automation: Designed an advanced automation script using Playwright for seamless interaction with web elements, ensuring efficient and reliable video uploads.
• This experience allowed me to master various automation tools and techniques, significantly improving my ability to develop and deploy efficient content management systems.