Looking for a job

Paarangat Rai Sharma

@paarangatraisharma

Message

Founding engineer building production-grade AI agent products end-to-end—fast, observable, and cost-optimized.

India

Message

What I'm looking for

I want a team where I can own end-to-end AI agent product delivery—tight latency, strong observability, versioned prompts, and measurable cost wins—building across web, mobile, and desktop with engineering rigor.

I’m a founding-engineer-grade AI product builder who ships production AI agent systems end-to-end. In 6 months, I shipped 8 products across production, beta, and pre-launch—covering iOS, web, desktop, and browser—while keeping quality high through observability, versioning, and cost/latency discipline.

At JettyAI, I built GutBut, a Claude-powered AI gut-health companion, from solo ownership through a TestFlight closed beta. I architected and delivered four codebases (~25k LOC TypeScript) with an agent designed around tool/function calling constraints, prompt workflows with diff-gated publish, SSE-based real-time streaming, and production LLM observability via PostHog—cutting nightly insight regeneration cost to near-zero using SHA-256 hash-dedup.

I’ve also built regulated, real-world healthcare AI with Docwise—shipping a GDPR/UKCA/ICO/DTAC/DSPT-aligned Chrome extension and clinician recording workflow with role-based personalization for NHS personas. Previously at NULLZEC, I engineered sub-3s multi-model LLM orchestration in an Electron desktop app using a low-latency pipeline with routing, circuit breakers, and self-healing performance modes—always prioritizing reliability under real constraints.

Experience

Work history, roles, and key accomplishments

Current

React Native Engineer (Intern)

Current

Valtrium

Mar 2026 - Present (3 months)

Owned end-to-end React Native delivery for 6 consumer iOS apps using Claude Code-driven agentic workflows, reaching 736+ downloads across live apps. Built and shipped products including Siddhartha and Kathavani (live on App Store) and delivered feature-complete builds/pending launches for Phoenix, WatchIQ, PokeScan, and a wake-up call app, plus internal analytics and content pipelines.

React Native Expo Claude Code Agentic Workflows TypeScript Hermes Streaming Support

Current

Full-Stack Developer (Intern)

Current

Docwise

Dec 2025 - Present (6 months)

Built an ambient AI scribe for NHS doctors across a Chrome extension, dashboard, marketing site, and NestJS backend, producing 178 commits in 10 weeks directly under the CTO. Implemented a Manifest V3 Chrome extension with patient consent flow and contextual fields for AI note generation, and delivered clinical recording UX with Canvas-based real-time audio visualization and browser navigation-los

TypeScript NestJS Chrome Extension Manifest V3 Offscreen Document API IndexedDB Canvas FFT Role Based Access Control Supabase

Early Team Member (Fullstack Engineer)

JettyAI

Mar 2026 - May 2026 (2 months)

Built GutBut, a Claude-powered AI gut-health companion, and shipped four TypeScript codebases (~25k LOC) including NestJS backend, Next.js web, Expo iOS app, and a superadmin for prompt management and LLM observability. Implemented an agent with tool/function calling and ephemeral caching to save ~400 input tokens per common call, and reduced nightly insight-regeneration cost to near-zero via SHA-

Prompt Engineering Tool Function Calling Claude Code TypeScript NestJS Next.js Prompt Versioning PostHog LLM Observability

Software Engineering Intern

NULLZEC

Jan 2026 - Feb 2026 (1 month)

Architected a real-time multi-model LLM orchestration pipeline in an Electron desktop app (screen capture → perceptual hashing → filtering → semantic dedup → streamed LLM output) achieving sub-3s end-to-end latency. Built the Electron application with frameless overlays, Zod-validated IPC across 40+ channels, and decomposed a 2,100-line orchestrator into 8 DI modules with 304 Vitest tests.

Electron DHash Perceptual Hashing Speculative Parallel Execution Circuit Breakers Adaptive Throttling Zod Validation Vitest

Education

Degrees, certifications, and relevant coursework

Bennett University

Bachelor of Technology, Computer Science (AI)

Grade: CGPA 9.54 / 10

Activities and societies: Dean's List; 3x hackathon winner.

Pursuing a B.Tech in Computer Science (AI) at Bennett University with a CGPA of 9.54/10 and Dean's List recognition. Awarded as a 3x hackathon winner.