# Suryanand: Solutions Architect & Founding Engineer

> Solutions architect and founding engineer who solves messy problems with software. About eight years taking products from zero to production: the Agent Serve AI workspace and Agent Live, plant-floor systems for Granules India, Syntegon, Hosokawa Micron, Sun Pharma and Azbil, freight software for Pegasus, the Plugzee EV app, and logistics at Google.

Last updated: 2026-10-21. Canonical URL: https://www.suryanand.com/

## Summary

- Name: Suryanand (Suryanand Sunil)
- Role: Founding Engineer at Araxis AI (2024 to present); previously engineer at Google, Logistics & Supply Chain (2020 to 2024); independent work in healthcare, e-commerce and internal tools (2017 to 2020)
- Based in: Pune, India and Dubai, UAE; works remotely
- Works as: solutions architect or founding engineer on a team, or on fixed-scope sprints
- Recent work: builds Agent Serve (AI workspace) and Agent Live (real-time sales-call copilot) with the team at Valnora & Elric (https://www.valnoraelric.com); led delivery of plant-floor and field systems for Granules India, Syntegon, Hosokawa Micron, Sun Pharma and Azbil; built Pegasus (freight) and Plugzee (EV super app)
- Studio: Valthrax (https://valthrax.com), open-source tools for LLM engineers under Apache-2.0
- Teaching: https://bootcamp.suryanand.com
- Contact: work@suryanand.com · https://calendly.com/work-suryanand/spark-revops-strategy-call

## About

I've spent about eight years moving products from the messy beginning to the part where they actually work: logistics at Google, healthcare automation and the core platform of an AI startup as its founding engineer. I care less about novelty and more about shipping the right thing: clear architecture, honest observability and code the next engineer can live with.

Clients and engagements: Alphabet (Logistics), Granules India (Pharma), Syntegon (Process equipment), Hosokawa Micron (Powder processing), Sun Pharma (Pharma), Azbil (Automation), Pegasus (Freight), Syngenta (Agritech), Hoag (Health), Relic Care (Health), Plugzee (EV), Ploy (Software), Aestlo (Commerce).

Skills: System design, Core platforms, TypeScript, Python, Next.js, FastAPI, Postgres, Temporal, LLM tooling, Voice AI.

## In numbers

- 8 years — shipping software that runs in production
- 13 companies — built for, from Alphabet to Sun Pharma
- 9 industries — from pharma lines and freight to fashion
- 4 systems — AR, 3D and digital-twin systems on the plant floor

## Selected work

### Agent Serve (Valnora & Elric · Platform architecture)

An AI workspace where teams build, deploy and manage AI agents. I architected the platform and lead its engineering, from multi-tenant data to the agent runtime.

- 6 modules: Chat, Workflows, Knowledge, Tables, Files, Logs
- 3 ways to build agents: visually, in chat or in code
- 1 workspace for agents, data and BI dashboards
- Tags: AI agents, Multi-tenant, Platform
- What changed: Teams build, run and govern AI agents in one workspace, connected to their business tools, instead of stitching together scripts, chat tools and dashboards.
- Capabilities: Chat for building in natural language, Visual workflow builder, Relays: AI coworkers with approvals, Knowledge bases and tables, BI dashboards, Multi-tenant access control, Run logs, block by block
- Case study page: https://www.suryanand.com/work/agent-serve
- More: https://www.valnoraelric.com

### Agent Live (Valnora & Elric · Real-time AI)

A real-time AI copilot for sales calls: live guidance while the call is happening, playbooks and compliance checks, and an admin portal for managers.

- 3 surfaces: desktop app, live service, admin portal
- Live guidance during the call, not after it
- 1 playbook library shared by the whole team
- Tags: Realtime, Voice, Desktop
- What changed: Reps get the next best question while the customer is still talking, and managers see playbook and compliance coverage across every call.
- Capabilities: Real-time call transcription, Live guidance prompts, Playbook library, Compliance checks, Desktop app, Admin portal with call metrics
- Case study page: https://www.suryanand.com/work/agent-live
- More: https://www.valnoraelric.com

### Plugzee (EV mobility · Product build)

An EV super app: find and reserve chargers, plan road trips with charging stops, book service and insurance, and pay from one wallet.

- 5 journeys in one app: charge, service, buy, insure, travel
- Live charger availability and pricing on the map
- 1 wallet and rewards across every journey
- Tags: Mobile app, Maps, Payments
- What changed: EV drivers plan, charge, service and insure from one app, instead of juggling a different app for every charger network and garage.
- Capabilities: Charger map with live availability, Charger reservation, Road-trip planner with charging stops, Service and test-drive booking, Insurance, Wallet and rewards, Partner command centres
- Case study page: https://www.suryanand.com/work/plugzee

### Granules India (Pharma · Forward deployed delivery)

Lyophilization process intelligence: every freeze-dryer, cycle phase, alarm and batch record on one live screen, with the audit trail on.

- 7 cycle phases tracked, from loading to complete
- 7 live readouts per lyophilizer
- 1 screen for every freeze-dryer on the block
- Tags: PLC + historian, Audit trail, 3D
- What changed: Operators and QA work from one live view of every freeze-dryer, with alarms, SOPs and batch records in the same place, instead of separate equipment screens and paper records.
- Capabilities: PLC and historian connectivity, Audit trail on every action, Recipe library, Alarm console with SOPs, Batch records and analytics, 3D equipment views, Role-based access
- Case study page: https://www.suryanand.com/work/granules
- More: https://www.valnoraelric.com/elric-works/granules

### Syntegon (Process equipment · Forward deployed delivery)

A fluid bed dryer digital twin: live sensor readings on a 3D model, drying curves, and an AI copilot that flags drift before it becomes a deviation.

- 3D live twin of the dryer
- AI copilot watching the drying curve
- AR view for engineers on the floor
- Tags: Digital twin, AI copilot, 3D
- What changed: Engineers see the dryer, its readings and the drying curve in one place, and the copilot points out drift while there is still time to act.
- Capabilities: Real-time 3D digital twin, Sensor overlays on the model, Temperature and airflow maps, AI process copilot, Recipes and work orders, Audit trail, AR view
- Case study page: https://www.suryanand.com/work/syntegon
- More: https://www.valnoraelric.com/elric-works/syntegon

### Hosokawa Micron (Powder processing · Forward deployed delivery)

Process and equipment intelligence across milling and drying: 3D equipment, live parameters, health scores, CIP/SIP monitoring and reports.

- 3D equipment viewer, plant down to the bearing
- Live parameters and equipment health scores
- CIP and SIP cleaning cycles monitored
- Tags: 3D viewer, Equipment health, Reports
- What changed: Operators and maintenance work from one model of the plant, moving from line to machine to live parameter without leaving the platform.
- Capabilities: 3D equipment viewer, Equipment hierarchy, Live parameters and KPIs, Equipment health scores, CIP / SIP monitoring, Process reports, Work order shortcuts
- Case study page: https://www.suryanand.com/work/hosokawa
- More: https://www.valnoraelric.com/elric-works/hosokawa

### Sun Pharma (Pharma field maintenance · Forward deployed delivery)

Field AI for on-site maintenance: smart glasses show the next step on the equipment, capture evidence and write the inspection report as technicians work.

- AR step guidance on the equipment itself
- Auto inspection reports when the job ends
- Live remote session view for maintenance leads
- Tags: Smart glasses, AR, Mobile
- What changed: Technicians work from the procedure on the equipment instead of from memory, and inspection reports are ready when the job is done.
- Capabilities: Smart-glasses pairing, AR step guidance, SOP-driven checklists, Photo, video and voice capture, Automatic inspection reports, Remote session view, Report library
- Case study page: https://www.suryanand.com/work/sunpharma
- More: https://www.valnoraelric.com/elric-works/sunpharma

### Azbil (Industrial automation · Forward deployed delivery)

On-site operations with AR: sign-in, pre-work compliance, guided execution, live data and evidence capture, ending in an audit-ready record, hands-free.

- AR hands-free field workflow on smart glasses
- Voice AI assistant for field engineers
- 2 languages in one interface
- Tags: AR, Voice AI, Compliance
- What changed: Field engineers follow one hands-free workflow from sign-in to sign-off, and every job ends with a complete, audit-ready record.
- Capabilities: Smart-glasses AR guidance, Pre-work compliance checks, Voice AI assistant, Live equipment data, Evidence and e-signatures, Automatic audit trail, Bilingual interface
- Case study page: https://www.suryanand.com/work/azbil
- More: https://www.valnoraelric.com/elric-works/azbil

### Pegasus (Freight · Product build)

A freight operating system: shipments, bookings, containers and exceptions in one command centre, with live tracking and AI insights on delay risk.

- 3 modes tracked live: ocean, air and road
- AI insights on port congestion and delay risk
- 1 command centre for every exception
- Tags: Web app, Live tracking, AI insights
- What changed: Operations teams see every shipment and every exception on one screen, instead of chasing carriers across email and spreadsheets.
- Capabilities: Command centre KPIs, Live multimodal tracking, Exception management, Bookings and rates, Document tracking, AI delay insights, Agent studio
- Case study page: https://www.suryanand.com/work/pegasus
- More: https://www.valnoraelric.com/elric-works/pegasus

## Open-source tools (Valthrax, Apache-2.0)

### Pressmark

Prompt version control and eval harness: SHA-addressed prompts, async runner, five built-in scorers, FastAPI + HTMX web UI, CI gate via --min-pass-rate.

- Install: `pip install pressmark`
- Site: https://pressmark.valthrax.com
- Source: https://github.com/Suryanandx/pressmark

### Switchboard

Terminal UI for OpenRouter: query 200+ models, switch live, stream markdown and track cost per message. Built on Textual + httpx SSE.

- Install: `pip install switchboard-tui`
- Site: https://switchboard.valthrax.com
- Source: https://github.com/Suryanandx/switchboard

### Murmur

Local-first agentic voice pipeline (mic → STT → LLM → TTS → speaker) with sentence-boundary buffering for sub-second perceived latency.

- Install: `pip install murmur-voice`
- Site: https://murmur.valthrax.com
- Source: https://github.com/Suryanandx/murmur

### Cue

Open-source Cluely alternative: a stealth desktop overlay invisible to screen-share, agent-driven answers, Ollama + OpenRouter routing and a scriptable HTTP daemon. Mac, Windows and Linux.

- Install: `pip install cue`
- Site: https://cue.valthrax.com
- Source: https://github.com/Suryanandx/cue

## Lab: side projects

- [Yoga LMS](https://yoga-lms.suryanand.com/): Course-first learning platform for yoga instructors and students (350+ users).
- [R2R Care](https://r2r-care.suryanand.com/): Patient onboarding and care workflows for a multi-clinic group (200+ users).
- [Fix My CSV](https://fix-csv.suryanand.com/): Repair messy CSVs in the browser, no login or tracking (500+ users).
- [MockCanvas AI](https://mock-canvas.vercel.app/): Prompt to API mockup, a weekend prototyping tool.
- [Task Manager](https://todoist.suryanand.com/): Minimal keyboard-driven task app (750+ users).
- [Threads, studied](https://threads-clone.suryanand.com/): Reverse-engineered feed, posts and real-time updates (350+ users).

## Services

### 1. Founding engineering

Your first technical partner. I own the platform end to end: architecture, the build and the first engineering hires.

- Includes: Core platform, Architecture, Observability, Hiring loop
- Terms: Ongoing (Full-time or contract)

### 2. Solutions architecture

I map the business problem to a system: integrations, data flows, security and the build plan, written down before anyone writes code.

- Includes: Discovery workshops, Reference architecture, Integration design, Security review
- Terms: 1–2 weeks (Fixed scope)

### 3. Forward deployed engineering

Embedded with the process owners, inside your IT and security rules, putting AI and data into production against a business target.

- Includes: On-site discovery, Systems integration, AI in production, Team handover
- Terms: 30-day sprint (Outcome-based)

### 4. 0→1 product builds

From a back-of-napkin idea to something people use in production, in about two weeks, without a rewrite later.

- Includes: Product scoping, MVP build, Deploy and monitoring, Handoff docs
- Terms: ~14-day sprint (Fixed scope)

## How a 0→1 sprint runs

1. **Discover** (Days 1–2): I sit with the people who'll use it and write down the one outcome that matters, with a number next to it.
2. **Architect** (Days 3–4): A one-page system sketch: data model, queues and failure modes. Boring where boring helps.
3. **Build** (Days 5–12): Daily deploys to a real URL. You click the real thing every evening, not a slide.
4. **Ship** (Days 13–14): Launch with observability on, docs written and code the next engineer can live with.

## Testimonials

> "Two weeks from Figma to a working app. We shipped to users on the call right after handoff."
> — Ananya Sharma, SaaS founder, Pune

> "Our junior devs picked up the code without one question. That alone was worth it."
> — Ishan Kapoor, CTO, PaymentFlow

> "Three other devs gave up on this build. Suryanand had it live in two weeks."
> — Priya Desai, Healthtech founder

> "We launched on day 12. Closed our seed round on day 19."
> — Ravi Mehta, Founder

## Teaching (https://bootcamp.suryanand.com)

- **Node, AWS and System Design** (Interview prep, 12 videos · 4 modules): Interview-ready Node.js, AWS fundamentals and cloud system design, built for backend engineers cracking offers.
- **System Design in Depth** (Architecture, 13 videos · 5 modules): Architecture layer by layer: clients to services to queues to databases, reasoning through real designs.
- **Database Architecture** (Data, 2 videos · 1 module): Sharding, read replicas and the patterns that show up at data scale. New modules as the syllabus grows.

## FAQ

### What kind of problems do you take on?

The messy ones where software, data and operations meet: AI platforms, real-time voice, plant-floor and field systems, and products that need to go from zero to production. I've solved these for pharma, process-equipment, freight and EV companies.

### How do we work together?

Whatever fits the problem: as a founding engineer or solutions architect on your team, or as a fixed-scope sprint. An MVP takes about 14 days; an architecture review one to two weeks.

### How fast can you ship an MVP?

Typically 12 to 14 days from kickoff to a live product with real users, built so it doesn't need a rewrite later.

### Which stack do you use?

Boring where boring helps: TypeScript, Next.js, Python, FastAPI, Postgres and Temporal. Picked for the problem, not the resume.

### Where are you based?

Pune and Dubai, working remotely with teams in any time zone.

### Can I pitch an open-source tool?

Yes, through valthrax.com/#pitch. If a few engineers would use it, it gets built and released under Apache-2.0.

## Contact

- Email: work@suryanand.com
- Book a 30-minute call: https://calendly.com/work-suryanand/spark-revops-strategy-call
- LinkedIn: https://www.linkedin.com/in/suryanand-sunil/
- GitHub: https://github.com/Suryanandx
- X / Twitter: https://x.com/suryanandkurup
- Pitch an open-source project: https://valthrax.com/#pitch