Summary
AI Platform Product Manager with 4.5+ years building agentic voice applications, LLM orchestration pipelines (STT → LLM → TTS), PromptOps platforms, evaluation frameworks, B2B SaaS and marketplace products. Combines deep engineering expertise with product execution to ship reliable, scalable applications with measurable business impact.
AI Voice Agents
LLM Orchestration
Agentic Workflows
PromptOps Platforms
B2B Marketplaces
P2P Finance Systems
Currently leading AI product strategy and execution at Vegrow across voice intelligence,
supply
& sales
intelligence, and finance systems.
Experience
Product Manager – AI & Automation · Vegrow
- Built and scaled an agentic voice platform (VG Voice) powering profiling, NPS, slot booking, and ops workflows; automated 30K+ calls/month, reduced cost per completed workflow by ~60% and increased scale 4×.
- Owned the full AI product loop: persona & prompt design → fallbacks → knowledge base → tool/function calling → post-processing → dashboards → retraining signals.
- Developed an API-first platform to integrate with multiple products (CRM, LMS, ERPs) across diverse use cases.
- Architected real-time STT → LLM → TTS orchestration with streaming and interruption handling; achieved <1.5s median response latency with clearer A/B-tested prompt variants, structured disambiguation, and interruption recovery, lifting workflow completion from 40% → 65%.
- Improved feedback outcomes: increased NPS from 30% → 65% using structured voice feedback, auto-summaries, sentiment tagging, issue categorisation, and streamlined reporting with closed-loop ownership to reduce churn.
- Defined AI-specific north stars: funnel visibility, completion rate, extraction accuracy, latency, cost/call, transfer-to-human, repeat-call rate and satisfaction rate.
- Integrated multi-agent evaluation into CI/CD as a release gate to prevent regressions in extraction quality, UX, and latency before deployment.
PromptOps Engine, Sales Intelligence & Integrations ·
Vegrow
- Built a centralized PromptOps system used across 5 GenAI applications (templates, variables, versioning, safe rollout mechanics); reduced iteration and deployment cycle time by ~2×.
- Launched AI-driven sales pitch intelligence to automate scoring and coaching; reduced manual QA effort by ~90% and improved conversion predictability.
- Shipped multiple business-critical integrations (telephony, finance, internal systems) with API-first design: authentication, webhooks/events, retries, queuing and observability.
Procure-to-Pay (P2P) – Business Finance Product · Vegrow
- Defined product scope for PR → PO → GRN → Invoice → Payments workflows; led a 5-day MVP build.
- Enabled ₹6 Cr+/month processing with automated workflows, reducing approval turnaround by ~40% and reducing financial leakage by ~7.5% via matching, validations and automated syncs to finance auditing tools.
Vegrow FxS – B2B Marketplace Product · Vegrow
- Strategized and led a B2B marketplace workflow product that reduced operational friction and enabled scale to ₹70 Cr+ GMV, improving end-to-end visibility and accelerating supplier payments.
- Deployed LLM-based OCR for handwritten slips (~70% accuracy) to auto-populate records, boosting speed and reducing errors during peak operations; improved field execution by ~2.5× while avoiding linear headcount growth.
Digital Quality Reports (DQR) · Vegrow
- Designed and launched a scalable quality assessment workflow product, enabling structured data capture and better customer-quality matching across 50+ SKUs; reduced operational effort by 1,000+ man-hours/month through standardized flows and role-based review.
- Increased defect traceability via standardized tagging and metadata schema, creating a high-quality dataset that now powers computer vision–based defect detection (pilot).
SDE-II (Tech Lead) · Vegrow
- Progressed from SDE-I to leading an 8-member team during Vegrow’s scale-up from ₹100 Cr → ₹500 Cr+ revenue, contributing to both core tech and team leadership.
- Led development of inventory traceability microservices using Golang, Kafka, MongoDB, and QR-based flows for perishable goods tracking at scale; mentored engineers and championed a quality-focused culture.
- Shipped event-driven systems with strong reliability patterns (idempotency, retries, DLQs, observability) to support hypergrowth.
- Built core platform services (auth/OTP, inventory state transitions, integrations, Computer Vision) and improved operational robustness through monitoring and incident hygiene.
Software Engineer · LivSYT
- Helped scale the project management SaaS from 0 → $0.5M annual revenue by building robust backend systems and manufacturing workflows serving customers across India, USA & Singapore.
- Designed event-driven notifications (AWS CloudWatch, Lambda, SES) to boost engagement and implemented RBAC/auditing in PostgreSQL to strengthen data integrity.
Computational Design Intern · Arth Design Build
- Worked in the R&D team on computational design automation tools using generative design and multi-objective optimization (genetic algorithms) to create data-driven layouts.
- Demonstrated client-facing prototypes showcasing computational design’s potential for scalable, efficient project planning.