Founding Engineer @ SlidelyAI · open to hard hardware + software problems

I build the whole stack — from the machine to the model.

Jai Shukla. I started at the bench fabricating biosensors, then kept building down and up the stack — firmware, precision machines, full-stack software, AI systems.

Origin: Biosensor scientist, 4 yrs → Now: AI & software engineer
Range: Rust firmware · motion control · React/TS · AWS · agentic AI
Based in Bengaluru, India · willing to relocate

See the builds ↓Download résumé Get in touch

FIG.01 — SVG TOOLPATH → MOTION · 100µm RESOLUTION

WP-01The through-line

Most engineers pick a layer and stay there. I go wherever the problem is.

If the tool I needed didn't exist, I built it — a potentiostat and its firmware, a 5-axis deposition machine, the control software that drives it, the AI that orchestrates it. The domains changed from electrochemistry to AI products; the instinct never did. Below is the same person solving problems at every level of the stack.

LAYER 00 — METAL

Machine & Firmware

Designed and built precision machines end-to-end. Rust firmware on RP2040 / ESP32 (embassy-rs), motion control, closed-loop vision, custom instruments.

LAYER 01 — SOFTWARE

Full-stack & Systems

React / TypeScript / Rust across desktop, web, and cloud. Real-time collaboration (Yjs/CRDT), microservices, AWS infra, cross-platform apps.

LAYER 02 — INTELLIGENCE

AI & Orchestration

Agentic workflows, RAG, computer-vision pipelines, context engineering — from independent video agents to secure multi-machine orchestration.

WP-02Featured builds

5-axis deposition machine — live run

CLIP 01 · motion + toolpath + deposition

Precision machineUtopic Tech · 2023–24

A 5-axis micro-deposition machine, built end to end

Role: mechanical design · firmware · control software · vision

Designed and built the machine from scratch — hardware, customized Marlin firmware, and the motion stack.

The hard part: killing tip vibration so material lands on a 100µm spot. I re-tuned the acceleration/deceleration profiles and paired high-count steppers with fine-pitch lead screws for ~50µm resolution, then closed the loop with image-based position correction to 10µm.

MarlinMotion controlStepper drivesOpenCVMechanical design

Potentiostat firmware — a CV / EIS sweep running

CLIP 02 · RP2040 · embassy-rs · live measurement

Embedded RustIIT Kharagpur · 2021–23

A self-built potentiostat and its Rust firmware

Role: hardware bring-up · async embedded firmware

Built firmware for a self-developed potentiostat on the RP2040 in Rust with embassy-rs, plus ESP32 firmware (esp-idf-rs) for real-time experiments.

Implemented the real electrochemical methods — CV, DPV, CA and EIS — across a 1µA to 100mA current-detection range: async control talking directly to analog front-ends, on a microcontroller.

Rustembassy-rsesp-idf-rsRP2040 / ESP32Analog / ADC

Rust control app + SVG→toolpath generation

CLIP 03 · egui · state machine · path planning

Rust desktop + toolingUtopic Tech · 2023–24

The software that drives the machine

Role: desktop app · path generation · Tauri mobile control

Wrote the machine's control software in Rust (egui), architected as a remote state machine so operators reconfigure and run new processes at runtime — no firmware changes.

Built an SVG-based process generator that compiles vector designs into multi-axis toolpaths, and a Tauri mobile app so a phone can drive the instrument directly.

RusteguiTauriState machinesPath planning

Biosensor fabrication + how the potentiostat works

CLIP 04 · NPTEL explainer · lab I set up

Fabrication & processNanobiosensor Lab, IIT KGP · 2021–25

Multiplexed biosensors, micropatterned at 50µm

Role: biosensor scientist · process development · lab build-out

Developed biosensor materials and my own fabrication process: a DLP printer to create 50µm patterns, used to build multiplexed sensors functionalized through electrochemistry.

Also rebuilt an RF/DC sputtering system from a full teardown and wrote its operating protocols, and set up the lab's measurement station from scratch. The clip includes an NPTEL segment where I explain how the potentiostat works.

DLP lithographyElectrochemistryRF/DC sputteringThin filmsML (SVM/NN/RF)

Real-time collaborative AI over Revit

CLIP 05 · Yjs/CRDT · multi-agent · custom query lang

AI systemsProject · 2026–present

Architectural Intelligence over Autodesk Revit

Role: systems architecture · agents · CAD integration

A multi-agent system with WebSockets and Yjs (CRDTs) for synchronized human–AI editing of Revit models, bridged to native CAD through a custom high-speed connector.

Authored a hybrid JSONPath–Emmet query language to traverse hierarchical Revit geometry for agents, plus a Qdrant-backed vector knowledge base and a long-context agent tool that chunks data to stop context pollution.

TypeScriptYjs / CRDTWebSocketsQdrantRevit / BIMAgents

AI slides, native inside PowerPoint

CLIP 06 · graphics tracing · cross-platform

Product · founding engineerSlidelyAI (YC) · 2025–present

Shipping an AI product across desktop, web & cloud

Role: Windows add-in · web app · AI orchestration · infra

Own the Windows software, the web app, the AI orchestration layer and the cloud infra. Unlocked web-app integration inside PowerPoint and ported legacy Windows add-ins to macOS with full parity.

Engineered a high-fidelity graphics-tracing pipeline (generative AI + computer vision + autonomous feedback agents) across distributed microservices; built OneDrive/SharePoint access via the Microsoft Graph API.

React / TSC# / VB.NETAWSGraph APIComputer visionMicroservices

WP-03The benchclick any image to enlarge

ADD IMAGE 01

015-axis deposition machine — the full rig I designed and built

ADD IMAGE 02

02Deposition head — tuned to land material on a 100µm spot

ADD IMAGE 03

03Custom potentiostat board (RP2040) running embassy-rs firmware

ADD IMAGE 04

04RF/DC sputtering chamber — rebuilt from a full teardown

ADD IMAGE 05

05Multiplexed sensor array, DLP-patterned at 50µm

ADD IMAGE 06

06Rust (egui) control software driving the machine

ADD IMAGE 07

07The measurement station I set up at the Nanobiosensor Lab

ADD IMAGE 08

08SVG toolpath → generated process, on screen

WP-04The pathbench → firmware → software → AI

FEB 2025 — PRESENT

Founding Engineer

SlidelyAI (Y Combinator-backed) · Bengaluru

Windows desktop software, web application, AI orchestration layer and cloud infrastructure for an AI-native slide product.

JUN 2024 — JAN 2025

Founder

Vasinya Yunaan · Pune

Built a platform analyzing 10,000+ patents for diagnostics market trends with RAG over OpenAI APIs; pitched investors and raised an initial $20,000.

MAR 2023 — MAY 2024

Scientist — Machine Design & Control

Utopic Tech Pvt. Ltd. · Pune

Designed and built the 5-axis deposition machine, its Rust control software, SVG toolpath generator, and vision-based position correction to 10µm.

MAR 2021 — 2025

Biosensor Scientist / Research Assistant

Nanobiosensor & BioDevices Lab, SMST, IIT Kharagpur

Four years developing biosensor materials, fabricating and patterning multiplexed sensors, building the potentiostat firmware, rebuilding a sputtering system, and ML for analyte prediction.

JUN 2021 — AUG 2021

Google Summer of Code Contributor

Open Chemistry, University of Pittsburgh

WebGL visualization pipelines for biomolecules and a UI for dynamic atom identification and labeling.

2016 — 2021

B.Sc. (Hons) + M.Sc., Chemistry

IIT Kharagpur

Master's thesis: k-means clustering for catalyst discovery in water splitting across 200+ journals.

WP-05The stack I build with

Languages

RustTypeScriptPythonCC# / VB (.NET)

Hardware & Firmware

embassy-rsesp-idf-rsRP2040 / ESP32Marlin / motionOpenCVPotentiostat designSputtering / DLP

Frontend & Apps

ReactNext.jsAngulareguiTauriWebGL

Backend & Real-time

Node.jsFastAPIActixHonoWebSocketsYjsSocket.ioPostgreSQL

Cloud & DevOps

AWSGCPAzureDockerKubernetesCI/CDGraph API

AI & Agents

Agentic workflowsRAGLangGraphQdrantContext engineeringComputer vision

WP-06Research & recognitionGoogle Scholar ↗

PUBLICATIONS · 5

Smartphone-integrated, reagent-free paper sensor for hematocrit measurement

Analytical Methods 15(29), 2023 · cited by 11 · device + image-processing server

Lab-on-chip electrochemical biosensor for rheumatoid arthritis

MEMS & Microfluidics in Healthcare, 2023 · cited by 6

Application of radiopharmaceuticals in diagnostics and therapy

Next-Gen Nanobiosensor Devices, 2022 · cited by 5

Electrochemical detection of cancer fingerprint: extracellular vesicle research from lab to market

Next-Gen Nanobiosensor Devices, 2022 · cited by 3

Role of biosensors in regenerative therapeutics

Regenerative Medicine: Emerging Techniques, 2023 · cited by 2

AWARDS

2nd — Siemens MakeItReal HackathonFull-stack security system + integrated hardware · 2019

1st — Product Design, INAE Youth ConclaveNovel bandage to control bleeding from severe wounds · 2019

3rd — Rural Technology HackathonDevice generating electricity from stove waste heat · 2018

Participant — Clinton Global Initiative UniversityFarm machine for stubble management · 2018