Prasad Patil - Solution Architect (Cloud · AI · Microservices) at TCS and PhD researcher at IIIT Nagpur, building the Dynamic Oracle Adaptation framework where quantum computing meets machine learning.
2026Notes are live ‐ a new writing section: the DOA framework, explainers, and build logs. Read the notes · RSS
2026HQ-DOA published ‐ “Quantum State Measurement System With Dynamic Oracle Adaptation”, IEEE Trans. Instrumentation & Measurement - first quantum-computing publication from IIIT Nagpur. DOI
2026FQ-DOA submitted to IEEE Trans. Emerging Topics in Computing - under review.
2026MindSpark released - zero-dependency collaborative mind-mapping. Repo
2024ACM India Anveshan Setu Fellowship - host mentorship at IIT Hyderabad.
Signal
The short version
I build systems that have to work - and lately, systems nobody has built before.
Live qubit · try itP(0) 50% · P(1) 50%
|ψ⟩
α|0⟩ + β|1⟩ - evolving
Adaptive oracles begin here: a state that hasn't decided. Press Measure and the superposition collapses to one outcome, weighted by the amplitudes at that instant. Re-cohere, and the wave returns.
For fifteen years at Tata Consultancy Services and TRDDC I've turned legacy enterprise estates into cloud-native architectures: COBOL and mainframe workloads migrated to Java and .NET microservices, compliance turned into a service, and GenAI put to work reading and rewriting code it was never trained to understand. Two granted patents came out of that work.
Somewhere along the way the questions stopped being about migration and started being about computation itself. As a PhD scholar at IIIT Nagpur I work where quantum computing and machine learning intersect: the DOA framework - Dynamic Oracle Adaptation - methods that let a quantum oracle adapt instead of staying fixed. The first of these became the first quantum-computing paper published from IIIT Nagpur.
Production code on weekdays, quantum circuits on weekends. I like it that way.
Every career is a trajectory through some latent space. Here's mine, plotted on the two axes that matter: classical → quantum and building → learning. Hover any point. The bottom-right quadrant is where this is all headed.
What if a quantum oracle didn't have to be fixed in advance? Instead of hard-coding the black box an algorithm queries, DOA methods synthesize it on the fly from what earlier measurements revealed.
In HQ-DOA that means composing two algorithms that are normally used apart: Simon's period detection uncovers the hidden structure of the problem, DOA turns that structure into a freshly-built oracle, and Grover's amplitude amplification then searches the space that structure has already cut in half - with no mid-circuit measurement to collapse the state along the way.
Schematic - in the hybrid DOA loop, measurement statistics drive a classical update that re-parameterizes the oracle between runs.
IEEE Transactions on Instrumentation & Measurement · 2026
Quantum State Measurement System With Dynamic Oracle Adaptation for Enhanced Parameter Estimation in NISQ Devices
P. P. Patil · J. Tembhurne · S. J. Choudhury · HQ-DOA
Integrates Simon’s period detection with Grover’s amplitude amplification through dynamic oracle adaptation. Eliminates mid-circuit state collapse by synthesizing measurement oracles from period-detection results ‐ lifting measurement success from 15% to 60% on the IBM Brisbane processor, at query complexity O((2n−1/k)½).
First quantum-computing publication from IIIT Nagpur
IEEE Transactions on Emerging Topics in Computing · 2026
Fully Coherent Dynamic Oracle Adaptation
P. Patil · J. Tembhurne · S. J. Choudhury
Under review
HQ-DOA pipeline · Simon → adapt → Grover
1 Simon ‐ find the hidden period
2 DOA ‐ synthesize the oracle
3 Grover ‐ amplify the reduced space
Each Simon query returns one linear constraint on the hidden period s. Once n−1 independent constraints pin it down, DOA rebuilds the oracle around that structure ‐ and because f(x) = f(x⊕s), every state pairs with a partner and the search space folds in half. Grover then runs over what's left, reaching the target in fewer queries than a blind search. Schematic illustration of the pipeline; the quantum steps are simulated classically, not a reproduction of the paper's hardware results.
Fixed vs adaptive oracle · same target, same start
Fixed oracle
Uƒ
phase θ = 180° ‐ every query, forever
6%θ 180°
Adaptive oracle
Uƒ(θ)
phase θ re-tuned from feedback each round
6%θ 180°
fixedadaptive0 queries
Both start identically. The fixed oracle keeps rotating past the target and tears its own amplitude back down ‐ it has no way to know it already arrived. The adaptive oracle reads the outcome each round, re-tunes θ, and locks on. Schematic illustration of the DOA idea ‐ a greedy feedback rule stands in for the adaptation strategy; not a reproduction of the paper's results.
Trajectory
Fifteen years of shipping
2023 - now
Solution Architect
TCS · MBU Apps Infra Engineering - AI.Cloud Unit, Nagpur
Leading Azure cloud-migration lifecycle tooling and large-scale modernization. Architected GenAI-driven COBOL→Java migration with contextual understanding of legacy systems; modular microservices via Domain-Driven Design.
2019 - 2023
Technical Architect
TCS · ISN Corporate Unit, Pune
Core contributor to Intelligent Services Networks - Spring Boot microservices and Azure integration; product architect for automating the software-delivery lifecycle.
2014 - 2019
Full-Stack Developer & Researcher
TRDDC · Ignio / Digitate Unit, Pune
Owned Compliance-as-a-Service end to end, research to product delivery - compliance engines and integrations across enterprise infrastructure.
2011 - 2014
Developer & Researcher
TRDDC · eTransform Unit, Pune
Java modules for data-center migration and transformation estimation - multi-threading, design patterns, performance-sensitive workloads.
