– Rebuilt the core data ingestion pipeline from the ground up, achieving a 9X speed up in end-to-end latency while implementing production-ready capabilities such as hierarchical clustering and knowledge graph generation

– Developed the data analytics pipeline for the company’s proprietary ModelMesh™ (autonomous reasoning engine) technology integrating an in-house Text-to-SQL model, thereby increasing accuracy by over 70%

– Led end-to-end development of the Network Topology Agent capable of processing billions of records every 10 days; owned production deployment on AWS EKS and designed GitOps-based CI/CD pipelines using ArgoCD, enabling reliable multi-environment releases at scale

– Closely partnered with AWS in launching the Network Topology Agent as a revenue-generating AWS

Marketplace offering; implemented cloud-agnostic user registration, subscription, and entitlement workflows (Featured on the company blog: https://articul8.ai/blog/smarter-gen-ai-agents-ready-to-deploy)

– Built a reusable AI agent orchestration architecture for an agentic compliance analysis product, reducing compliance analysis turnaround time by 99% for customers across regulated industries

– Introduced production alerting to the platform by identifying and eliminating root causes of failures; reduced mean time to detect and respond (MTTD/MTTR) from hours to minutes, dramatically improving on-call team efficiency

– Built and maintained high-performance systems using Python, Ray (distributed job orchestration), ArangoDB, Milvus, GraphQL APIs, OpenTelemetry for distributed tracing, and related observability tooling

– Partnered closely with Customer Success and Solutions teams; regularly joined customer calls to demo

product capabilities, gather feedback, and translate real-world use cases into product and engineering improvements

– Implemented production-grade CLI commands to validate data integrity and runtime health across distributed services, significantly reducing incident triage time

– Orchestrated the observability stack for AWS EKS clusters using the Grafana-Prometheus-Loki stack and integrated alerting with Microsoft Teams using AlertManager

– Set up Application Load Balancers and created monitoring dashboards to ensure uptime of showcase

applications for clients

– Developed a financial-domain LLM benchmark to evaluate model performance across quantitative accuracy and latency, enabling data-driven tradeoffs in production model selection

• Served as a GSR in the Department of Neurobiology and Behavior, School of Biological Sciences in UC Irvine for the Winter 2024 quarter

• Fine-tuning Large Language Models (LLMs) using use-case specific data

• Developing LLM-powered applications using LangChain to implement Retrieval Augmented Generation (RAG) for question-answering and code-generation tasks

• Developed a sandbox environment to run code generated by leveraging LLMs

• Optimized on quality and performance by customising prompt templates and document retrievers

• Spearheaded automated monitoring tasks for proactive failure reporting of critical processes on production servers

• Attained a return offer for a full-time role due to exceptional performance and innovative contributions

• Pioneered the development of an advanced Android Application for comprehensive data collection from farmers in Andhra Pradesh, implementing a well-structured SDLC approach.

• Instituted industry-standard modules and robust security measures, including Salting and Hashing, to fortify the application against potential Man-in-The-Middle attacks and tested private APIs with Postman, ensuring seamless integration and enhanced software functionality.