About

Hi, I’m Swati Modi, a Software Engineer 👩‍💻, passionate about products and a maker at heart 🚀

I like figuring things out 🕵️‍♀️, Side projects are my love language, and there is usually something half-working on my laptop that I am weirdly excited about 😬.

Work Link to heading

Full-time Roles Link to heading

  • Last9 (Levitate TSDB) Scaled a real-time time-series warehouse processing over 1 trillion samples/month. Built ingestion pipelines, streaming aggregations, and billing systems. last9.io | Levitate

  • Delhivery (OS1 Logistics SaaS) Built distributed backend pipelines for order processing and dispatch management in a large-scale logistics platform. delhivery.com | OS1

  • Fynd (Glamar Virtual Try-On SDK) Developed a real-time virtual makeup and face filters SDK for beauty and retail apps. fynd.com | glamar.io

Consulting Experience — Founded Helix Software Services 👩‍💼 Link to heading

At Helix Software Services, I lead a focused engineering team that partners with startups and growth-stage companies to build, optimize, and ship reliable systems—especially where speed, scalability, and real-world performance matter.

My core expertise lies in backend systems and distributed infrastructure, with hands-on work across low-latency, high-throughput platforms and practical machine learning deployments.

🛠️ What We’ve Worked On

  • Backend Scaling for a Unified Search Platform Optimized and stabilized the backend for a personal search engine indexing cloud drives, emails, and bookmarks. Reduced infrastructure costs and improved query response times by redesigning distributed services, enhancing caching layers, and refining indexing and retrieval strategies.

  • Real-time Communication Backend for VOIP & Chat Designed and implemented backend services for a real-time VOIP and messaging solution, handling call signaling, session management, and low-latency message delivery. Focused on scalability, fault tolerance, and seamless mobile connectivity for users.

  • Real-time On-Device Background Removal & Pose Tracking Engineered real-time, on-device systems for background segmentation and fitness pose correction using MediaPipe, optimized to run at 30 FPS on mid-range mobile devices with no server dependency. Integrated into video and fitness apps for efficient offline processing.

  • Hand Tracking Pipeline Latency Optimization Diagnosed and resolved latency spikes in MediaPipe’s iOS hand tracking pipeline, significantly improving responsiveness and stability for production gesture-based applications.

What we don’t do 🙈🙉🙊 Link to heading

  • Backend systems that look shiny in docs but cry under load 😭
  • “AI solutions” that need a supercomputer to say hello 🥵
  • Research that lives its best life… in a PDF 🎭

We are hands-on builders who care more about what works than what sounds fancy. Clean infra, fast code, and getting things to run where they actually need to run, whether it’s shaving latency, squeezing models onto devices, or making backend systems cheaper and more reliable.

Tech I have worked-on/specialize-in or find interesting so far 🤓 🧐 Link to heading

  • Distributed systems
  • On-device ML
  • WebAssembly and AR

I strongly believe in end-to-end productivity and having a holistic product vision than just ticking feature boxes.

I like reading books, both fiction and non-fiction 💫 🕵️‍♀️ 📚, and writing about tech @swati-modi and blog.

I enjoy running, biking, lifting things, cooking healthy, and pretending to meditate. 🏃‍♀️ 👩‍🍳 🧘‍♀️