Jun 2025 – Aug 2025 · 2 mos · Washington DC-Baltimore Area · On-site

• Advanced Strategy One’s production LLM-based Text-to-SQL system by introducing schema-grounded multi-step reasoning, focused schema retrieval, and execution-guided validation/repair, improving SQL execution accuracy from 82% → 86% and reducing invalid/hallucinated SQL by 25% across ~12K daily queries spanning 4.5K tables and 120+ schemas.

Jan 2025 – Jun 2025 · 5 mos · Washington DC-Baltimore Area · On-site

• As part of the CSL lab at virginia tech worked with Booz Allen in Partnership with CCI. Pioneering the future of 5G security with AI-driven autonomous testing! Harnessing advanced machine learning to predict, detect, and neutralize vulnerabilities in telecom networks—eliminating manual security testing and revolutionizing real-time threat mitigation. Pushing the boundaries of AI in cybersecurity to set new standards for O-RAN and shape the next era of secure, self-defending 5G networks!

Nov 2024 – Feb 2025 · 3 mos · Washington DC-Baltimore Area · Remote

• Engineered and trained a LLM to autonomously map Suricata rules, transforming fragmented threat signatures into an intelligent, self-evolving knowledge base of MITRE techniques. This system didn’t just detect threats—it learned, reasoned, and predicted attack patterns, bridging rule-based security with AI-driven cyber intuition. The result? A smarter, context-aware defense layer that redefined how security teams interact with intrusion detection whilst saving costs and optimising inference-time!

Jul 2022 – Aug 2024 · 2 yrs 1 mo

• As a front office analyst within the Low Touch Cash Equity Analytics team at Wells Fargo, I specialize in the research and development of quantitative analytics models and trading strategies for the U.S. equities market. My work involves a rigorous analysis of time-series data, leveraging both current and historical trading patterns to construct robust predictive models. These models are instrumental in providing actionable insights for traders, guiding buy/sell decisions and optimizing stock portfolio performance.

Jan 2022 – Mar 2022 · 2 mos · Hybrid

• Participating undergraduate research student in the Google Research Week featuring research lectures and discussions on ML techniques, addressing their developments, limitations, and critiques.

Jan 2022 – Jan 2022 · 0 mo · On-site

• My work centered around creating secure, compliant, and effective engagement tools for financial institutions, which included Text, Video, Audio, Secure Chat, and Social Messaging. In my position, I also focused on the integration of our platform with essential IT systems such as Symitar and Corelation. Employing advanced analytics, I derived insights into customer engagement, which guided the refinement of our tools to boost operational efficiency, client interaction, and staff productivity.

Jun 2021 – Dec 2021 · 6 mos · Hybrid

• At Samsung, I reimagined how users discover content by building a next-gen recommendation system for the Discover app. I designed the server-side architecture, harnessing Java, Kotlin, AWS, and DynamoDB to power real-time content delivery and intelligent data retrieval. By integrating Knowledge Graphs and embedding generation, I transformed personalization, making recommendations smarter and more adaptive. To push the boundaries further, I developed Data-Augmented Deep Learning models, implemented privacy-preserving AI, and optimized event-driven processing with Apache Kafka—driving a 15% surge in user engagement. This work didn’t just refine recommendations; it redefined how users interact with content in a dynamic, AI-driven ecosystem.

Jun 2020 – Feb 2021 · 8 mos · Pilani, Rajasthan, India

• Part of the top 5 startups selected from a pool of 200+ startups for startup incubation in BITS Pilani

May 2020 – Jun 2021 · 1 yr 1 mo · Mumbai, Maharashtra, India

• At Expound, I played a pivotal role in scaling an in-house SaaS product, driving $3,000 in revenue with an annual projection surpassing $10,000+ across 10+ clients. I engineered an AI-powered E-Commerce platform for an imitation jewelry retailer, optimizing customer experience for a $6,000 annual turnover business. By leveraging React.js, Node.js, and Postgres, I delivered data-driven software solutions, integrating AI-powered analytics to enhance decision-making. To maximize profitability, I developed churn prediction models, increasing SaaS revenue by $5K, and implemented AI-driven recommendation algorithms, boosting E-Commerce retention by 12%—transforming data into growth.

Apr 2020 – Jun 2020 · 2 mos · Vadodara, Gujarat, India · Remote

• Engineered a MATLAB-based power analysis model to optimize load sharing across industrial systems, ensuring seamless power distribution through parallel generating sources. Worked on Smart Grid Technology and validated Generation Control Systems using a real-time hardware simulator to enhance system efficiency. Designed a power management system to mitigate overloading and stability issues, increasing power reliability, cost efficiency, and redundancy, ensuring uninterrupted operations even during outages.

Apr 2019 – Jul 2019 · 3 mos · Mumbai Metropolitan Region

• Developed an AI-powered Sentiment Analysis tool to enhance customer support automation, streamlining ticket workflows from generation to closure. Increased classification accuracy and automation by 15%, leading to a 2% boost in customer satisfaction and a 5% improvement in operational performance. Leveraged Google & Twitter APIs with NLP to detect fake profiles, flag negative tweets, and predict user engagement patterns, improving user credibility by 1.2% and growing the brand's social media following by 600 users.

Jan 2019 – Jan 2020 · 1 yr · Mumbai, Maharashtra, India

• Founded Servofast, a B2BC consumer electronics startup designed to streamline the value chain by aggregating key stakeholders and enhancing operational efficiency.

Jan 2019 – Jan 2019 · 0 mo · Mumbai, Maharashtra, India · On-site

• Built a responsive e-commerce platform, enabling seamless online ordering with integrated payment gateways. Developed an admin portal with role-based access controls, ensuring secure and efficient user management. Successfully deployed the platform integrated with Google Analytics and SEO optimization, improving performance and accessibility.