Devansh Jain

Voice is already the next big thing! Let's make it right & affordable together. Join us!

Largely a Platform Role - Engineering the Infrastructure that runs AI backbone for Rippling. - Part of a high functioning 30 member AI Engineering working across US/India/Europe.

Largely a Product Role. - Took on a challenging IC role, within Rippling's third highest vertical (250M$ ARR), working with teams across US & India in PST/IST. - Built Sandbox Environment to experiment large scale (100M$+) changes/rollouts. - Built Flow Framework, on top of Temporal -- simplifying chained orchestration across 50K+ object changes. - Helped simplifying Operational Rigor.

📍 Project Iota – Field Execution Infra • Embedded directly with Market Managers across multiple cities — posed as one of them to understand: → Workflow bottlenecks → Repeated hacks → Behavioral edge cases and incentive gaming • 🛠️ Built two foundational systems: – 🗺️ Beat Framework: territory assignment, geo-tracked logins/logoffs – 📌 Task Framework: goal-driven prompts, offline caching, and integration-ready structure • Designed backend and mobile architecture for low-connectivity ops, fast retries, and multi-region scale → Tools became the default for market expansion and agent accountability 📚 Catalog Platform – Quality Bridge in Re-Arch • Embedded into senior Catalog re-architecture group (Principal Architect, GM, senior SDE3s) • Owned “Quality of Catalog” charter during the transition from legacy to new platform → Focused on classification, SKU grouping, vendor override patterns, and taxonomy consistency • Acted as translator between product, ops, and new infra → Ensured new platform met real-world data expectations → Prevented silent breaks across pricing, search, and inventory tools 👥 Execution Scope • Operated as founding engineer for Market Ops stack — with direct trust from field leadership • Prioritized delivery speed, platform cleanliness, and resilience in messy real-world contexts • Balanced zero-to-one engineering with long-term maintainability patterns 🧰 Stack: Node.js, Golang, Redis, PostgreSQL, Android clients, internal orchestration tools 📈 Impact: Enabled scalable field ops for offline GTM expansion; ensured catalog data quality survived platform overhaul without regressions

💬 Assist Platform (Essentials Vertical) • 🧠 Operated as Lead Product Engineer (SDE-IV) on Assist v2 — powering chat, voice, ticketing, and resolution flows for 400+ agents • Scoped and shipped a redesigned agent dashboard focused on speed, context, and resolution flow → Unified chat/call/ticket surfaces → Cut agent cognitive load → Reduced bounce rate on issue raise flows, improved AHT consistency • Owned Assist for Essentials (Food, FMCG, Medicine) verticals (~90% of business revenue) • Mentored SDE-III handling Non-Essentials, ensured tooling reuse + convergence across verticals ☎️ Telephony Re-Architecture • Partnered with Ziyad (remote SIP architect) to rewrite udaan’s Telephony stack → Migrated from Twilio to a modular, in-house SIP-based system → Focused on stateless architecture, call traceability, and failure recovery → Achieved ~85% infra cost reduction, increased call success rate to ~99% • Owned high-level product-architecture bridge: agent feedback → call flows → system re-design → Made call context + failure explainability part of Assist UI → Gave CX leaders visibility into where resolution broke down (and why) 👥 People & Trust Scope • Led pod execution post-layoff — with only senior engineers retained • Mentored directly toward platform independence (Gautam → SDE-IV track) • Trusted to converge Assist v2 across verticals, align category PMs, and handle delivery from ground up 🧰 Stack: Node.js, Golang, SIP/Asterisk, Kafka, Redis, internal orchestration 📈 Impact: Unified Assist platform org-wide, rewrote telephony for long-term control + cost efficiency, rebuilt core CX systems with minimal headcount

🧱 Platform Systems Built • 🧩 Designed and shipped udaan’s Discovery Platform from scratch → Replaced static layouts with real-time merchandising infrastructure → Enabled non-tech teams to launch personalized storefronts in minutes using: – Tile Builder – Page Builder – Campaign Scheduler → Became the default interface for all homepage, seasonal, and vertical-led campaigns • 🛍️ Led platform work on udaan Studio → Built for seasonal brands (Adidas, Puma, Reebok) to run range-based pre-sales → Supported pre-orders, long-tail inventory discovery, and buyer-side planning • ⚡ Architected backend for Quick Commerce (Fresh/FMCG) → Focused on regional personalization, availability-aware browsing, and bundling logic → Handled catalog overrides, pricing tiers, and GTM-aware discovery flows 👥 Team & Execution Ownership • 👨‍👩‍👧‍👦 Led a cross-functional pod of engineers and designer(s); owned roadmap + delivery • 🧠 Mentored engineers who later led org-critical systems (Search, Payments, Coupons, Onboarding) • 🔄 Designed for internal reuse and experimentation: infra served Ops, Merch, Category, and Studio teams • 📦 Created platform artifacts that became standard: – Dynamic layout engines – Real-time campaign rendering – Multi-vertical content targeting 🧠 Org-Level Influence • 🏗️ Operated as a Staff-level IC + EM hybrid, focused on platform leverage • 🤝 Embedded with stakeholders across merchandising, ops, and product — shaped roadmap from ground-level ops feedback • 🪴 Trusted by Head of Engineering to scale internal tooling across categories while keeping infra cost sustainable 🧰 Stack: Node.js, React (internal), Redis, Solr, internal CMS + campaign engines 📈 Impact: Reduced campaign iteration from days → minutes, improved discovery-led conversion, and laid the foundation for Assist 2.0 personalization

🛠️ Systems Designed & Owned • ☎️ Built udaan’s first in-house Telephony stack → Replaced Twilio/Exotel with SIP-based infra → Cut costs by ~80%, improved call reliability to >98%, and enabled full call traceability • 🧩 Architected Assist v1 primitives: ticketing, chat, IVR routing, agent skill trees → Enabled 400+ agents to operate with contextual resolution flows and in-dashboard actions • ⚡ Designed fast, agent-focused workflows → Unified chat/call UI, AHT-reducing raise/resolve actions, and cross-channel continuity → Improved Average Handle Time (AHT) by ~20% and increased agent utilization • 📉 De-risked org-wide support dependency on vendors → Migrated from Freshdesk + Exotel to internal tooling stack in under 12 months • 📊 Built observability dashboards for SLA tracking, agent throughput, and call failure rates → Gave ops full visibility + tunability without engineering intervention 🧠 Strategic Ownership • 🤝 Worked across Product, Ops, and Infra to align latency, cost, and scale tradeoffs • 🏗️ Defined core patterns for how udaan scaled CX across verticals (grocery, pharma, lifestyle) • 🪜 Laid groundwork for future personalization, campaign-led CX, and self-serve resolution 🧰 Stack: SIP/Asterisk, Node.js, WebRTC, MySQL, Redis, Kafka, Freshdesk APIs 📈 Impact: Built systems powering CS during udaan’s hypergrowth — when support was a core revenue lever

> Worked with SpiceJet to development E2E SpiceJet App for Airplane Cockpit in Xamarin.Forms and Azure > Maintained a team of 4 developers as the Scrum Master

-> Analysing Sentiment Patterns on Texts using IBM Watson Engine

I completed UnderGrad course in Computer Science and Engineering. Major Subjects studied during the duration are : Microprocessor & Micro controllers Object Oriented Analysis and Design Object Oriented Programming Theory of Computation System Software & Compiler Design