Professional Summary
Artificial Intelligence & Data Science engineer with hands-on production experience building real-time conversational voice AI systems at Pype AI, from internship through a full-time engineering role. Currently building fusion-runtime, an open-source voice AI runtime competing with LiveKit and Pipecat. Strong foundation in voice AI architecture, observability tooling and full-stack development.
Experience
Pype AI (Singularity Corp Pvt. Ltd.), Bengaluru
Jan 2026 – Sep 2026
AI Intern (Jan–Jun) → AI Engineer — Conversational & Voice AI, Full-Time (Jul–Sep)
- Built and then owned Whispey, an observability platform and AI voice analytics dashboard for monitoring production voice AI systems, from initial build through full-time ownership.
- Engineered and maintained real-time conversational agents using LiveKit and Pipecat, automating patient interaction workflows at production scale.
- Audited production call data — latency bottlenecks, call issues and conversational edge cases — to improve the reliability of live AI voice pipelines.
Projects
Self-hosted voice AI runtime · Apache-2.0 · Python
- Built a single-process STT→LLM→TTS voice agent runtime that streams between stages instead of serialising through network hops, so synthesis begins while the LLM is still generating — cutting perceived response time. Measured ~490 ms turn-end-to-first-audio on an RTX 3090 with a 7B GGUF model at 127 tok/s.
- Designed a runtime-adapter contract keyed on inference engine (
ctranslate2, llama_cpp, onnx, openai_http), letting any Whisper checkpoint or GGUF model run with zero new code, enforced by a reusable conformance test kit.
- Wrote acoustic echo cancellation from scratch (STFT filter bank, NLMS adaptive FIR, GCC-PHAT delay estimation), validated at 30 dB suppression on laptop speakers, enabling server-side barge-in with sub-10 ms client-side playback flush.
- Built a deadline-ordered, concurrency-limited scheduler sharing one loaded model across sessions; load-tested and publicly documented the degradation curve (4 callers: ~740 ms, 8 callers: ~4.6 s).
- Shipped v0.1.0 to PyPI with a Typer CLI, a browser client over AudioWorklet + WebSockets, Prometheus metrics and CI across Python 3.11–3.13.
FaMo (Famous Model) App
React, Supabase, Vercel
- Developing a comprehensive social media platform tailored for the modelling industry, connecting talent with agencies.
- Architected a real-time backend using Supabase to handle user profiles, social feeds and secure authentication.
- Deployed the responsive React frontend on Vercel, configuring a custom domain managed via Hostinger.
Technical Skills
- Languages: Python (async/await), TypeScript, Java, JavaScript (AudioWorklet), SQL, Bash
- Voice AI & Real-Time Systems: LiveKit, Pipecat, WebRTC, WebSockets (bidirectional PCM streaming), STT/LLM/TTS pipeline design, barge-in & turn detection
- AI/ML Inference: faster-whisper/CTranslate2, llama-cpp-python (GGUF), ONNX Runtime, Kokoro-82M, Silero VAD, PyTorch/torchaudio, Hugging Face Hub, Gemini 3 Pro, OpenCV, YOLOv11
- Web Development: FastAPI, Uvicorn, Pydantic v2, Next.js, React.js, Node.js (MERN), Supabase, HTML5, CSS3
- Observability & Tooling: Prometheus, structured telemetry, GitHub Actions, Docker, PyPI packaging, Typer CLI
- Cloud & Platform: Vercel, Google Cloud (GCP), Cloudflare Pages, Git/GitHub, Power BI
Education
CMR Institute of Technology, Bengaluru
2022 – 2026
B.E. in Artificial Intelligence & Data Science
Relevant coursework: Data Structures & Algorithms, DBMS, Operating Systems, Artificial Intelligence.
Achievements & Certifications
- Kaggle: ranked 2271/13131 in "Titanic — Machine Learning from Disaster".
- Open Source: author of fusion-runtime, published to PyPI under Apache-2.0.
- Infosys Springboard: completed certifications in Deep Learning, NLP and Data Science.
Interests
Technical: building with LLMs, quantitative aptitude Personal: motorcycle touring (Himalayan 450), learning languages
Tip: print this page (Cmd/Ctrl + P) to save a PDF copy.