Headline

Senior Software Engineer | AI/ML Systems & Full-Stack | LLM Infra, MCP, RAG | Apache Contributor | 6 YOE

About

Senior software engineer with 6+ years designing and shipping production-grade AI systems, full-stack platforms, and real-time data infrastructure at enterprise scale across fintech, e-commerce, and cloud product companies. I specialize in LLM/GenAI application development: RAG pipelines, MCP-based agentic systems, multimodal inference optimization, and prompt engineering for domain-specific automation. At Oracle, I built Maui Coding AI, an MCP + RAG system using Claude Opus and vector embeddings that reduced frontend development effort by 80% and accelerated feature delivery 5x. I also architected a Visual IR Framework that cut multimodal context window usage by 60%, reducing enterprise inference costs by 42%. On the full-stack side, I've led end-to-end product builds in React, Node.js and TypeScript from a conversational banking AI (Chanakya) on legacy fintech infrastructure, to a gold loan digitization platform that scaled the loan book to ₹19,000 Cr. I design for scale: streaming APIs, Microservices, Webhook architectures, and PII/SPDI-compliant data pipelines. I'm also an Apache Fluss Committer, contributing to the open-source streaming Lakehouse that unifies real-time and historical analytics for AI workloads. I'm open to Staff / Senior SWE roles in AI infrastructure, LLM tooling, or full stack product engineering particularly at teams building AI first platforms.

Location

Bengaluru

AI stack

claude code

codex

cursor

langchain

langgraph

opencode

Work experience

Senior Member of Technical Staff at Oracle, Apr 2025 - May 2026 · 1 yr 2 mos, Bengaluru, Karnataka, India, • Tasked with accelerating frontend development velocity across product teams working on Oracle's Maui UI framework, designed and shipped Maui Coding AI, an MCP-based agentic code generation system using Claude Opus, RAG pipelines, and domain-specific vector embeddings, reduced frontend development effort by 80% and accelerated feature delivery 5x across 3 product teams. • Enterprise LLM inference costs were unsustainable due to oversized multimodal context windows, architected a Visual IR Framework to extract structured context from images and video for AI inference, using optimized multimodal preprocessing and context compression, reduced context window size by 60% and cut inference costs by 42%, improving latency for real-time AI workloads.

Lead Software Development Engineer at Freecharge, Mar 2023 - Aug 2024 · 1 yr 6 mos, Mumbai, Maharashtra, India, • Legacy fintech infrastructure was creating bottlenecks in customer servicing with no conversational AI layer, led system design and full-stack development of Chanakya, a conversational banking AI integrating LLM-powered NLP on existing core banking systems, with streaming API encryption and LLM-based PII/SPDI redaction ensuring regulatory compliance and reduced customer servicing time by 30–50% while maintaining 100% PII/SPDI compliance at real-time throughput. • Gold loan operations were entirely manual, causing slow disbursals and poor loan book growth, led end-to-end product development of the Midas Gold Loan Platform from scoping, architecture, and delivery to integrating AI based valuation APIs, secure payment flows, and core banking system connectors, digitized the gold loan vertical and scaled loan book value to ₹19,000 Cr.

Senior Frontend Developer at Cordelia Cruises, Jun 2021 - Mar 2023 · 1 yr 10 mos, Bangalore Urban, Karnataka, India, • Static pricing was leaving revenue on the table during peak demand periods, designed and implemented demand-optimization algorithms for dynamic pricing, integrated with a webhook based microservices architecture for real-time fare updates across booking channels, improved revenue capture during high-demand periods and reduced time-to-market for pricing changes. • Siloed teams and waterfall delivery cycles were slowing feature releases; drove adoption of Agile workflows, CI/CD pipelines, and cross-functional sprint planning across engineering and product, reduced time-to-market by 35% and cut customer journey completion time by 55%.

Software Developer at Infinite Potential, Nov 2019 - May 2021 · 1 yr 7 mos, India, • A Siemens IIMS legacy system was creating costly inventory overhead and slab tracking errors in steel production; replaced the system with a custom infrastructure management platform and built a computer vision pipeline for automated slab indexing and inventory tracking; reduced inventory overhead by 38% and eliminated manual slab identification errors.

Co-Founder at urbanstop, Nov 2018 - Nov 2019 · 1 yr 1 mo, Greater Bengaluru Area, • Identified a gap in hyperlocal e-commerce for urban consumers with no existing technical platform; co-founded Urbanstop and built the full e-commerce platform from scratch — owning architecture, frontend, backend, vendor integrations, and go-to-market execution end-to-end; successfully launched and operated the platform, gaining hands-on experience across the full product and business lifecycle.

Education

Bachelor's degree, Computer Science, Visvesvaraya Technological University

Bachelor of Engineering, Computational Science, Visvesvaraya Technological University

Public links

github: https://github.com/niknegi