Vishnu Vardhan Reddy Yeruva is an Enterprise Generative AI Lead, SAP BTP Architect, and software engineer who has successfully bridged the gap between heavy enterprise architecture and mobile-edge consumer applications. As the creator of Sahayi AI, an iOS application that uses generative AI to help couples stay emotionally connected, he has combined advanced prompt engineering, procedural audio synthesis, real-time partner synchronization, and health tracking into a single, cohesive platform.
What sets Yeruva’s work apart is his conviction that generative AI should go beyond enterprise productivity. In this TechBullion interview, he shares how he applies his background in enterprise cloud architecture to the technical backend of Sahayi, his vision for the future of AI-powered companionship, and why the most powerful applications of generative AI are the ones that serve the deeply human need for emotional connection.
1) Please tell us a little more about yourself and your background
My name is Vishnu Vardhan Reddy Yeruva, and I am a software engineer and AI architect based in New Jersey. I hold a Master’s degree in Health Informatics from Rutgers University. Professionally, I work as an SAP Functional Consultant and Generative AI Lead for Mygo Consulting Inc., actively building and deploying enterprise AI solutions for major clients such as Johnson & Johnson. My day-to-day work involves architecting complex, large-scale systems using SAP Business Technology Platform (BTP), Python, and React to solve massive corporate data challenges.
However, Sahayi AI started as a deeply personal project. I was in a relationship and wanted to build something that could act as a caring, continuous presence for my girlfriend when I couldn’t physically be there. I decided to take the advanced LLM orchestration techniques I use in the enterprise space and apply them to a consumer mobile application. What began as a personal gift quickly grew into a full-featured iOS application that uses generative AI to bridge emotional distance. Today, Sahayi is live on the App Store, and I’m exploring ways to expand its capabilities to help couples everywhere.
2) What is Sahayi AI, and what problem does it solve?
Sahayi AI is an AI-powered companion app designed to help people stay emotionally connected with their loved ones. At its core, it addresses the emotional gap that forms when life gets busy, when partners are physically apart, or when someone simply needs a reminder that they are cared for.
The app uses OpenAI’s GPT-4o models to generate deeply personalized content—morning love notes, emotional support letters, relationship horoscopes, and real-time conversational chat—all tailored to the specific relationship dynamics of the user. But Sahayi goes far beyond just AI chat. Drawing on my background in health informatics, it integrates health tracking for water intake, meals, and medication, featuring real-time synchronization to a partner’s device through.
Firebase. It includes a virtual companion pet whose mood reflects the user’s daily health habits and a sleep mode with procedurally generated soothing sounds.
Generative AI has largely been applied to code generation and enterprise retrieval. Sahayi demonstrates that one of the most powerful use cases for AI is making people feel cared for.
3) From a technical standpoint, how does Sahayi use generative AI differently from a standard chatbot?
The key difference is in the prompt architecture and context window management. Standard chatbots use static system prompts with generic instructions. Sahayi builds its AI personality dynamically from user-configured relationship data.
During onboarding, users provide relationship metadata—whether they are spouses, parents, siblings, or friends. This feeds directly into a dynamic system prompt that shapes every API interaction. The AI responds as though it were created by the user as a gift of love for their specific partner, creating an emotional resonance that generic chatbots cannot achieve.
We also separate our API call pipelines to optimize latency and context. Conversational chat maintains a rolling context window of the last ten messages, while our “Open When” letters (for specific emotional states like sadness or stress) use isolated, zero-shot API calls with meticulously engineered prompt templates. For example, our “Open When You’re Sad” prompt includes eight specific directives about acknowledging pain with empathy and building confidence through specific affirmations.
4) You mentioned procedural audio synthesis for the sleep feature. Can you walk us through how that works technically?
This was one of the more interesting mobile engineering challenges. Rather than bloating the app size by shipping large audio files, we built a procedural sound engine using Apple’s AVAudioEngine framework that generates these sounds algorithmically in real time.
We implemented three distinct noise generation algorithms. White noise is generated through random sample values between -1 and 1. Pink noise, used for rain and forest sounds, is generated using the Voss-McCartney algorithm, which produces noise with equal energy per octave. Brownian noise, used for ocean and fireplace, is generated through a random walk with a leaky integrator to prevent signal drift and clipping.
Each generated noise buffer is then routed through an effects chain—an equalizer for frequency shaping and a reverb unit for spatial depth. Rain, for instance, passes through an aggressive low-pass filter at 800 Hz with a large room reverb to create the muffled, immersive quality of rainfall.
5) The app includes a virtual pet whose mood is tied to health habits. What was the design thinking behind that?
This was inspired by the psychology of Tamagotchi-style engagement, adapted for a health and wellness context. The virtual pet—a dog named Whistle—serves as a gamification layer that encourages users to maintain healthy daily habits without the app feeling like a clinical health tracker.
The pet has three core stats: hunger, thirst, and happiness. When a user logs a meal or water, the respective stats refill. These stats naturally decay over time, creating a gentle, recurring incentive to log health metrics. The pet’s mood is computed from the average of all three stats and transitions through six visual states. The key insight is that people who might ignore a sterile push notification will actively engage with a virtual companion that is visibly affected by their behavior.
6) How does the partner synchronization system work, and what role does Firebase play in the architecture?
Partner synchronization is built on Firebase Cloud Messaging and Firestore. When a user completes a health-related action, the app triggers a real-time push notification to their partner’s device. This transforms individual health tracking into a shared caregiving experience.
There is also an intelligent medication adherence check. If a user enables the “notify partner if meds are missed” setting, the app schedules a daily local notification. If the user hasn’t logged their medication by that time, a reminder fires, and the partner is alerted. Authentication is handled entirely through Sign In with Apple, providing a secure, privacy-respecting way to link partner accounts without exposing sensitive credentials.
7) Where do you see generative AI heading in the personal companionship space over the next three to five years?
I believe we’re at the very beginning of a fundamental shift. Over the next few years, I expect to see AI companions that maintain longitudinal emotional memory—genuine understanding of a relationship’s history, growth, and patterns over months or years, utilizing advanced Retrieval Augmented Generation (RAG) applied to behavioral data.
We will see AI that can detect emotional distress through behavioral signals—changes in app usage patterns or typing cadence—and proactively offer support. The companies that will lead this space are the ones that treat emotional AI not as a novelty, but as a core infrastructure layer for human wellbeing.
8) What has the user response been so far, and what features are you planning next?
The response has validated our core thesis: people want technology that feels human. Users consistently tell us that the personalized “Open When” letters and the morning love notes resonate deeply. We’ve heard from long-distance couples and partners caring for loved ones with chronic conditions who rely on the app’s gentle health synchronization.
On the product roadmap, I am exploring deeper integrations with Apple HealthKit to pull in passively collected health data, such as sleep duration, enabling the AI to generate even more contextually aware check-ins. We are also prototyping voice interaction pipelines so the AI companion can feel more present through natural, low-latency speech. Ultimately, our north star is an AI that anticipates what both partners need and brings them closer together.
9) What advice would you give to developers wanting to build AI for emotional or relational use cases?
First, start with a real human problem, not a technology demo. I built Sahayi because I was personally experiencing the pain of being physically apart from someone I love. That authenticity shapes every design decision.
Second, treat prompt engineering as a rigorous software design discipline. The difference between a generic AI response and one that makes someone feel genuinely understood comes down to the specificity embedded in your backend prompts.
Third, respect privacy absolutely. Build on privacy-respecting infrastructure, minimize data collection, and be transparent. Trust is the currency of emotional technology.
10) Looking ahead, what is your long-term vision for Sahayi AI and generative AI in daily life?
My long-term vision is that generative AI becomes an invisible, caring layer in people’s relationships—something that quietly supports connection and emotional well-being without demanding constant screen time.
Sahayi started as one person’s gift to the person they love most. My vision is to scale that exact feeling of being cared for, making it available to every couple and family navigating the distances of modern life. Generative AI has the power to be the most empathetic technology ever created; we just have to choose to build it that way.
Learn More:
- Download Sahayi AI: Available on the Apple App Store
- Connect with Vishnu Vardhan Reddy Yeruva: LinkedIn Profile
