PineBrook Technologies
Founder's OfficeBuilt and shipped complete digital platforms: issafoundation.co.in, pinebrooktechnologies.com digital presence, and specialsprings.com India market platform.
India · Product & AI
28.6139° N, 77.2090° E
I turn ambiguous problems
into shipped
AI products.
I work across product discovery, AI systems architecture, rapid prototyping, and production delivery, building AI products that solve real problems.
Founder's Office @ PineBrook ·Sub-5ms Inference @ GetCred ·ICCCNet 2026 Edge VLM Research
-60%
Unqualified Demos
PineBrook Technologies: shortened sales cycles by 3 weeks using interactive workflows.
<5ms
Tree Inference
GetCred: 2M+ transactions with zero raw statement retention under DPDP Act 2023.
28
Indian States
Medikabazaar: automated multi-state payouts, cutting turnaround from 3 days to under 10 minutes.
ICCCNet
2026 Manchester
Hardware-aware edge visual-language system research and the Cache Cliff framework.
Founder's Office · 2026
PineBrook digital platforms & workflow automation
Delivered web platforms (issafoundation.co.in, pinebrooktechnologies.com, specialsprings.com) and built an interactive demo & compliance automation system.Digital platformsProduct discoveryPrototypes
Shipped 3 production platforms, cut unqualified enterprise demos by 60%, shortened sales cycles by three weeks, and reduced compliance documentation effort by 80%.
Fintech Product · 2025
GetCred alternative credit intelligence
Explainable cashflow underwriting and alternative credit scoring for India’s 80M+ gig and informal workers.Product discoveryXGBoost & SHAPUPI Telemetry
Enabled instant 0–900 GigScore evaluation, sub-5ms decision latency on 2M+ simulated monthly transactions, and cryptographically auditable zero-retention reports.
Operations systems · 2025
Medikabazaar operations data
Moved a payout workflow from multi-day manual handling to an operational system teams could trust.Data pipelinesOperations designReliability
Payout turnaround dropped from three days to under ten minutes across 400+ monthly transactions; 10,000+ records across 28 states supported a 35% reliability improvement downstream.
Systems research · 2026
Cache Cliff edge-AI research
A hardware-aware way to choose vision encoders for edge visual-language systems.BenchmarkingEdge inferenceLatency analysis
Introduced the Cache Cliff terminology to make non-linear hardware trade-offs explicit; readers should validate recommendations on their target hardware.Where the work happened
From product discovery to production systems.
Research Fellow
Edge AI systems researchBenchmarked eight vision architectures across four hardware platforms for edge visual-language workloads.
GetCred
Product & systems builderDesigned an alternative credit scoring engine (0–900 GigScore) for 80M+ unbanked gig and informal workers using UPI behavioral signals.
Medikabazaar
Data & operations engineerReduced payout turnaround from three days to under ten minutes across 400+ monthly transactions.
Technical Architecture
Sub-5ms inference and measured operational gains.
Product Discovery & Compliance
PRD AuthoringSolution ScopingInteractive DemosDPDP Act 2023Rapid MVPsTelemetry & Funnels
AI Systems & Edge Latency
Edge VLM BenchmarkingSub-5ms Tree InferenceXGBoost & SHAPTensorRT & TFLiteCache Cliff FramingVLLM Optimization
Full-Stack Distributed Systems
TypeScript / ESNextNext.js 16 (App Router)Cloudflare Workers & D1Neon PostgreSQL 17Drizzle ORMWebSocket / VoIP Audio
ICCCNet 2026 PresentationMay 2026
The Cache Cliff: Why Edge AI Latency Isn't Linear
A hardware-aware framework benchmarking 8 vision architectures across 4 platforms. Explains non-linear memory and cache bottlenecks without overstating theoretical FLOPS.
Read Systems PaperFrequently asked questions
Direct answers on architecture, research, and delivery.
What do you specialize in?
I specialize in product discovery, AI systems architecture, edge AI latency optimization, sub-5ms tree inference pipelines, and enterprise workflow automation.
What production AI systems have you built?
I served in the Founder's Office at PineBrook Technologies (Jun-Nov 2026), building issafoundation.co.in, pinebrooktechnologies.com digital presence, and specialsprings.com India market platform while cutting sales cycles by 3 weeks and compliance docs by 80%; built GetCred alternative credit intelligence (sub-5ms zero-retention tree inference across 2M+ transactions); and Medikabazaar operations data pipelines (reducing payout turnaround from 3 days to under 10 minutes).
What research have you published or presented?
I presented hardware-aware edge VLM research and introduced the Cache Cliff phenomenon at the International Conference on Computing and Communication Networks (ICCCNet 2026), benchmarking 8 vision architectures across 4 edge platforms.
What is the Cache Cliff in edge AI?
The Cache Cliff is my hardware-aware analytical framing for interpreting the non-linear latency spikes that emerge when vision-language models exceed on-chip cache or memory bandwidth limits on constrained edge hardware.
How do you handle data compliance under DPDP Act 2023?
For GetCred, I designed an ephemeral feature extraction architecture that performs sub-5ms inference on UPI telemetry and discards raw bank statements immediately, achieving zero statement retention compliant with India’s DPDP Act 2023.
How can teams or companies collaborate with you?
You can connect with me via email at mail@yashvardhan.dev, book a discovery call on Cal.com (cal.com/yashvardhans/coffee), or reach out on LinkedIn (linkedin.com/in/yashvardhan04).
Let's make the hard part tractable