Professional Experience

I have worked across AI, fintech, healthcare, geospatial systems, life sciences, video technology, and agritech. My focus is on building dependable systems, guiding teams with clarity, and shipping products that solve real problems. Below is a concise account of my experience, written to highlight scope, decisions, and measurable outcomes.

• Relocated an organization’s headquarters while maintaining operational continuity. Led the formation of new technical and non-technical teams, instituted lightweight processes, and aligned stakeholders on priorities. The transition improved cost structure and shortened delivery cycles without compromising quality.
• Built an internal platform that reduced SaaS product development timelines from weeks to hours. Introduced standardized templates, guardrails, and automated scaffolding so teams could ship production-grade services quickly and with confidence. The platform enabled faster iteration, clearer ownership, and more reliable releases.
• Directed the rollout of multiple AI-native products in taxation, hiring, and venture validation. Coordinated product, engineering, and GTM functions to ensure the right problems were solved first. Emphasized clear specifications, sensible defaults, and steady incremental releases to reach adoption quickly.
• Designed systems to analyze millions of regulatory filings at scale. Developed rule-based parsing, event detectors, and resilient monitoring with end-to-end latencies as low as 10–20 milliseconds. Focused on correctness, fault tolerance, and transparent observability to support high-stakes workflows.
• Built real-time audio streaming servers using gRPC, WebSocket, and HTTP/2 for ASR applications. Extended base models with domain- and language-specific training to improve accuracy for Indian regional languages. Ensured secure streaming, consistent throughput, and graceful degradation under load.
• Delivered a geospatial extraction system achieving approximately 5 cm mapping accuracy from panoramic imagery. Contributed to national-scale digitization initiatives, producing clean, structured street-level asset data. Filed multiple patents in geospatial analytics and GIS to protect core methods.
• Authored lightweight motion detection algorithms used in video applications for children. Optimized for client-side constraints and real-time responsiveness. Balanced compute cost with accuracy to ensure stable experiences on modest devices.
• Built data annotation pipelines for 2D images and 3D LiDAR. Combined server-side orchestration with client-side tools to improve throughput and quality. Introduced QA loops, sampling, and metrics so teams could trust labels at scale.
• Led migration from legacy SVN to modern Git-based workflows. Consolidated repositories, introduced CI/CD, and formalized code review practices. Mentored teams through the transition and established practical standards that improved clarity, release cadence, and overall reliability.
• Prototyped a Bluetooth Low Energy validation tool that parsed OTA logs and rendered sequence diagrams. The prototype matured into a widely used utility, reducing manual analysis time and improving issue triage.
• Co-founded an agritech initiative using computer vision and edge ML to deter wild animal intrusions in farmlands. Deployed field-ready systems, validated impact, and secured public funding. Balanced robustness with cost to ensure sustainability in real conditions.
• Applied document analytics to life sciences by extracting bioassay signals from patent texts and diagrams. Built image-processing pipelines that outperformed heavier deep learning baselines in speed and reliability. Delivered results that supported research and investment evaluation.
• Contributed to interactive data collection tools with gesture recognition and motion capture directly in the browser. Used lightweight ML and careful optimization to deliver responsive experiences without specialized hardware.

I value precision, steady progress, and clear communication. My aim is to build systems and teams that remain dependable well after the initial launch.