Building Humbl Voice AI @ QWR

Prashant Singh

Voice AI · Backends · Realtime systems

I build production systems that talk, literally: low-latency AI voice agents, telecom integrations, and the scalable Python/Django backends behind them. 4+ years shipping at QWR, JAXL, and Infosys.

Portrait of Prashant Singh

Prashant Singh

Software Engineer

Question What's Real (QWR)

Current Focus Building Humbl Voice AI
Hardware Integration QWR Glasses (Wearables)
Core Specialization Realtime Audio · WebSockets · Low Latency
Primary Stack Python · Django · Flutter · WebSockets
Audio Latency Benchmark ~350ms TTFA
  • 4+yrsexperience building & shipping production software
  • 3companiesWorked at: QWR · JAXL · Infosys
  • ~350msTTFA (time to first audio) achieved on an AI voice agent

Work Experience

  1. Jul 2026 — Present

    Software Engineer @ QWR

    • Flutter
    • Python
    • Django Channels
    • Gemini Live API
    • Building Humbl, QWR's AI voice assistant mobile app in Flutter, delivering natural, low-latency conversational experiences.
    • Integrating Humbl with QWR Glasses to bring voice-driven assistance to wearable hardware.
  2. Aug 2024 — Jun 2026

    Software Engineer @ JAXL

    • Python
    • Django
    • DRF
    • PostgreSQL
    • Twilio
    • Elasticsearch
    • GCP
    • Architected a high-scale Call Campaign Engine supporting Manual, IVR, and AI-driven conversational calling, with decoupled agent execution so each module scales independently.
    • Integrated Plivo for inbound/outbound calling and Twilio for voice verification and OTP-based SMS authentication.
    • Resolved PostgreSQL index bloat on the latest_activity field: traced index churn from per-activity updates and redesigned the update strategy to restore query performance.
    • Built a call-tag feature letting agents annotate and full-text search calls, reducing average call-retrieval time.
    • Implemented the Shopify integration flow: authentication, store selection, and HMAC validation.
  3. Aug 2022 — Aug 2024

    Software Engineer @ Infosys

    • Python
    • Flask
    • RAG
    • Milvus
    • Angular
    • GCP
    • Built a Retrieval-Augmented Generation (RAG) API with Flask and Milvus, enabling semantic search over large enterprise document corpora.
    • Contributed to dataset preparation and fine-tuning workflows for LLM projects: data cleaning, prompt engineering, and evaluation.
    • Angular developer on a Smart City demonstration project.
    • Deployed and managed applications on Google Cloud Platform.
  4. Apr 2022 — Jul 2022

    Software Engineering Intern @ JAXL

    Backend internship; rejoined JAXL full-time in 2024.

    • Python
    • Django
    • Apple Pay
    • Raised test coverage from 60% to 85% with model constraints and unit tests.
    • Handled Apple in-app subscription lifecycle notifications (renew, expire, cancel, subscribe) server-side.

Selected Projects

Vargi-Bot

3D warehouse simulation for the e-Yantra Robotics Competition: a robotic arm sorts packages by medical priority using ROS Manipulation and OpenCV, with MQTT + Google Apps Script syncing orders to a live dashboard.

  • ROS
  • Python
  • OpenCV
  • MQTT
  • JavaScript

Construct o Bot

Two-wheeled robot with line following, wall following, path planning, and pick-and-place, including a power circuit fabricated on PCB.

  • Embedded C
  • ATmega2560
  • V-REP

Technical Expertise

Languages

  • Python
  • JavaScript
  • Dart
  • C/C++
  • HTML/CSS

Frameworks

  • Django
  • Django REST Framework
  • Django Channels
  • Flask
  • Flutter
  • Angular

Databases

  • PostgreSQL
  • Supabase
  • MongoDB
  • Redis
  • Milvus
  • Elasticsearch

AI & LLMs

  • Gemini Live API
  • Claude
  • RAG
  • Function calling
  • AWS SageMaker

Telecom & Realtime

  • Twilio
  • Plivo
  • Exotel
  • WebSockets
  • MQTT

Cloud & Other

  • GCP
  • ROS
  • OpenCV
  • REST APIs

Education

B.Tech, Electronics & Communication Engineering Institute of Engineering and Technology, Lucknow · 2018 – 2022 · GPA 8.0

Let's Connect

What's next? Something big, I hope, and I can't wait.

I'm genuinely itching to take on the next hard problem. Voice agents that feel human, backends that hold up under real traffic, realtime systems where every millisecond counts: this is the stuff I can't stop thinking about, and I want to go deeper.

I'm actively looking for software engineering roles in backend, voice AI, and realtime systems, with a team that ships ambitious things and pushes me to grow. If you're building something exciting, I'd love to hear about it. My inbox is always open, and I reply fast.

Say hello