Clinical generation & evaluation
Multi-pass documentation, structured diagnoses and plans, source-supported capture, and provider review.
SENIOR AI ENGINEER · FORWARD DEPLOYED
Intelligence,
in production.
I’m Parth Ghayal. I build GenAI, agent systems, and data platforms—and bring them into the workflows people depend on.
From problem discovery and architecture to evaluation, integration, and customer rollout. Ownership across the full path to production.
Enterprise AIAgents & automationData platforms
Building in healthcare? Explore my domain experience SYSTEM / EXPLORERInteractive architecture
04
Connect retrieval, analysis, and summarization to produce useful insights.
Evidence → reasoning → reviewClick a node to inspect
Explore the engineeringARCHITECTURE → IMPLEMENTATION → EVALUATION → DEPLOYMENT
50+
Business-tool integrations
Source coverage of the enterprise platform
8+
Customer practices
Direct engineering delivery in healthcare
2
Merged contributions to Ray
Distributed AI infrastructure
01 ENGINEERING RANGE
Across the stack.
Accountable for the outcome.
AI engineering, data depth, and the customer context to bring a system into daily use.
REASONING THAT CONNECTS TO ACTION
Design the behavior. Engineer the system.
I build retrieval, structured generation, and agent workflows around the task: the context a model needs, the actions it can take, and how its output will be evaluated.
- Separate retrieval, reasoning, and output validation.
- Ground actions in observed state and known facts.
- Use human review where a workflow needs judgment.
In practice: executive intelligence, a conversational learning assistant, clinical generation, and a local-model browser agent.
02 SELECTED ENGINEERING WORK
Different domains. Systems thinking.
Enterprise intelligence, production infrastructure, healthcare, and agents. Explore the scope I owned and the decisions behind each system.
5 projects
03 FORWARD DEPLOYED ENGINEERING
From the first conversation
to the next release.
I stay close to the customer and the code. This is the delivery loop behind my work.
Find the constraint behind the request.
Work directly with users to understand the workflow, the failure, and the result they need.
WHAT LEAVES THIS STAGE
A concrete problem and acceptance criteria
04 DOMAIN DEPTH / HEALTHCARE
AI engineering.
Clinical fluency.
I bring hands-on experience with the details healthcare teams care about: source-supported notes, diagnosis capture, lab data, provider review, and EHR write-back that respects each practice’s workflow.
8+
Nephrology practices
Direct engineering point of contact
~95%
Note accuracy
Reported in internal evaluations
How the note result was evaluated
Internal evaluations compared generated notes with transcripts and included provider review. The reported improvement was from 36% to approximately 95%; documentation-time comparisons reported approximately 90% less time. The sample size and detailed rubric are not published.
Documents, labs & retrieval
Document ingestion, OCR and lab extraction workflows, pgvector retrieval, and normalization review APIs.
EHR integration & write-back
Veradigm, TouchWorks, and eClinicalWorks / Redox workflows, with practice-specific controls for notes, orders, and statuses.
Operational controls
Tenant-scoped APIs, PHI-conscious logging, idempotent writes, and account-level switches for model and workflow changes.
05 INTERACTIVE SYSTEMS
Inspect the thinking behind the output.
Explore how context, constraints, and human review shape a system’s behavior.
ILLUSTRATIVE DEMO
SOURCE SNAPSHOT
Monthly recurring revenue$100k → $125k
Operating spend$70k → $77k
Planned delivery items40
Completed items32
Illustrative company data. Values are fixed for this demonstration.
EXECUTIVE PERSPECTIVE
Growth & delivery
Growth is up. Delivery needs attention.
Monthly recurring revenue increased 25%, from $100k to $125k.
The team closed 32 of 40 planned delivery items (80%).
Read revenue and execution together before setting the next target.
06 EXPERIENCE
Engineering depth.
Product perspective.
From data science and AI product development to enterprise intelligence and customer-facing production systems.
Full resumeSenior AI Software Developer
Nephrolytics AI
Clinical AI, structured documentation, EHR integrations, and direct engineering delivery for 8+ nephrology practices.
Senior AI & Data Engineer
SQOR.ai
AI insights across 50+ integrated business tools, multi-agent workflows, and executive and investor reporting.
AI Engineer / AI Product Manager
Radical AI
An educational AI assistant for students exploring AI, GenAI, LLMs, and machine learning. Progressed from engineering to product responsibilities.
07 RESEARCH & OPEN SOURCE
Work that extends beyond the product.
Published research, deep-learning investigations, and merged contributions to distributed AI tooling.
PUBLISHED PAPER / 2021
GAN-based image cartoonization
Exploring surface, structure, and texture representations for transforming real images into a cartoon-style format.
Read the paperPublication title
“CONVERSATION OF REAL IMAGES INTO CARTOONIZE IMAGE FORMAT USING GENERATIVE ADVERSARIAL NETWORK” — title as printed.
ACADEMIC RESEARCH PROJECT
System Identification with Neural ODEs
Comparing neural ODEs and liquid time-constant networks on three-body dynamics, with attention to performance and computational trade-offs.
Explore the report08 ENGINEERING PERSPECTIVES
Architecture, shared.
Writing about the systems, trade-offs, and practical work behind applied AI.
01
Agents & infrastructure
MCP in the Enterprise
The protocol, the trust boundary, and the engineering that still belongs to you.
Aug 2026Read on Medium
02System design
Architecting a Real-Time Multi-Tenant AI Analytics Engine
A technical exploration of the architecture behind an enterprise AI analytics platform.
Feb 2026Read on Medium
03Shipping AI
From Prototype to Production
What changes when an AI system moves beyond the notebook and into daily use.
Jan 2026Read on Medium
09 THE PERSON BEHIND THE WORK
A journey made possible by someone else’s belief.
Before the models, the degrees, and the deployments, there was my mother.
A single parent and a fashion designer running her own boutique, she worked to give me an education—from English-medium school and IIT entrance coaching to my master’s at Boston University.
Her support made this journey possible. My interest in AI began with hands-on learning at PICT, grew through research and teaching at BU, and continues in the systems I build with customers today.
Boston UniversityMaster’s in Artificial IntelligenceTeaching assistant · BA810
Pune Institute of Computer TechnologyB.E. Computer Engineering · 2018–2022AI/ML honors · Years 3 & 4
PLACES THAT SHAPED ME
Where it began
Aurangabad
My mother raised me as a single parent after my father passed away. A fashion designer with her own boutique, she often worked more than eighteen hours a day to support my education.
10 INDUSTRY & COMMUNITY
In the conversations shaping AI.
APR28–302026
ATTENDEE · SAN FRANCISCO
Stripe Sessions
At Moscone West, exploring the intersection of AI and business infrastructure—including Sam Altman’s conversation with Patrick Collison.
Explore the conversationMAR112026
ATTENDEE · NEW YORK
MCPconference
A day at Loft39 focused on the Model Context Protocol, agent frameworks, enterprise integrations, and production infrastructure.
Explore the conferenceLET’S BUILD SOMETHING USEFUL
A real problem?
That’s where I start.
For senior applied AI and forward deployed roles,
complex integrations, and production AI challenges.