I build the backend systems that let large companies control and account for business travel. At KAYAK I designed the approvals and trip reporting frameworks behind KAYAK for Business. Outside work I write about trustworthy AI, and review for conferences and technical publishers.
2024 to present · architect and sole backend engineer
A corporate travel platform is only usable by a large employer if it can enforce the employer's travel policy. The approval system I replaced blocked travellers before they could book, and when I analysed its production data it turned out to be failing as a control: very few requests were ever declined, and many were never acted on at all.
I designed the replacement. Travellers book normally, the approval request is created and routed at checkout, and a declined booking is cancelled automatically inside the airline's free cancellation window. Administrators choose what happens when an approver does not respond, so no booking is left pending indefinitely.
Launched in February 2025 on a programme of about ten people. Both earlier approval mechanisms were retired, and this is now the only approval method in the product.
Later projects by other engineers and teams, including multi-level approvals and backup approvers, extend the design rather than replace it.
In 2026 I added an engine that checks bookings made by human travel agents against the employer's policy. Those bookings had bypassed policy entirely, a gap that had been open for three years.
Companies need every booking tagged with their own accounting and compliance data: a cost centre, a project code, a reason for travel, a reason for booking outside policy. KAYAK used to meet each of these requests with custom code for one customer at a time, and a single field could take months.
I designed one general framework instead. A company's administrators define the fields they need, travellers answer them at checkout, and the answers travel with the booking into the company's own reporting and finance systems.
New reporting requirements are now configuration, set up by the customer's administrators without engineering work.
The framework has grown from six standard field types to twelve, and other teams have built post-booking editing and event reporting on the same data model.
Other recent work at KAYAK
Rail distribution through SilverRail for Amtrak and UK rail, covering exchanges, split ticketing, refunds and order reconciliation; 3-D Secure payment authentication across providers; and incident response for checkout across global distribution systems and rail suppliers.
Publications
Book chapters
Introduction to Trustworthy AI.
A. A. Potdar. Chapter 1 in Trustworthy AI Systems: Engineering Secure, Scalable, and Responsible Intelligence for Real Applications. Springer Nature, 2026.
The Evolution of Cloud Computing: From Virtual Machines to Serverless.
A. A. Potdar and P. Goswami. Chapter 2 in Next-Generation Technologies in Cloud Computing: From AI and Security to Sustainability. Wiley, 2026.
Evolution, Challenges and AI's Role in Cloud Engineering.
A. A. Potdar and S. Sudarsan. Chapter 1 in Cloud Engineering with AI for Achieving Scalability, Reliability, and Cost Efficiency. Springer Nature, forthcoming.
ISBN 978-3-032-19415-2
CI/CD and AI-Accelerated Quality.
S. Sudarsan and A. A. Potdar. Chapter 6 in Cloud Engineering with AI for Achieving Scalability, Reliability, and Cost Efficiency. Springer Nature, forthcoming.
ISBN 978-3-032-19415-2
Conference papers
A Multi-Agent AI Framework for Proactive Insider Threat Detection and Data Breach Containment in Enterprise Networks.
N. S. P. K. Yadati, L. H. Akkireddy, V. K. Polurouthu and A. A. Potdar. EAMCON 2026, Bangkok. Taylor & Francis proceedings.
Best Paper Award
Earlier
3D Character Generation from Images Using Convolutional Neural Networks.International Journal of Scientific Research in Computer Science, Engineering and Information Technology, 2019.
Digital Game-Based Learning: Design and Development of an Educational Game.International Journal of Research and Analytical Reviews, 2019.
Writing and open source
2026
HackerNoon. Articles on the cost and safety of AI-generated code, agent infrastructure, and distributed systems on cloud platforms. Ranked HackerNoon's top writer in AI/ML and in Programming in June 2026.
Loshbot. An open-source, local-first document assistant in Java 21 and Spring AI, built to learn retrieval, guardrails and private model hosting end to end.
Review and service
2026–27
Artifact Evaluation Committee, EuroSys 2027, the European Conference on Computer Systems.
2026
Programme committee reviewer, US-RSE 2026, the US Research Software Engineer conference.
More than 100 paper reviews for about twenty conferences, including IEEE URTC 2025 at MIT, IEEE iGET 2026, IEEE MIPR 2026, IEEE INDISCON 2026, IEEE InCODE 2026 and the 8th IEEE ICAC 2026.
2026
Technical reviewer for books. Manning: Building LLM Applications with DSPy (credited in the acknowledgements) and Architecting for Autonomy. AVA (Orange Education): Ultimate Agentic AI for Event-Driven Architecture (credited as technical reviewer). Chapter reviews for Springer Nature and Wiley.
2026
Session chair: IEEE TIC 2026 (June), ICSIST 2026 (August), and IEEE GICITE 2026 at Technische Universität Berlin (November).
2025
Application reader, Young Leaders of the Americas Initiative (YLAI) 2026 Fellowship, a U.S. Department of State programme.
Talks
Aug 2026
Invited speaker, ICSIST 2026, International Conference on Secure Information Systems and Technologies. Rayat Bahra University, Mohali, India.
Jan 2021
Chief guest, IT Colloquium 2021, St. Francis Institute of Technology, Mumbai.
Experience
2026–now
Senior Software Engineer, KAYAK (Booking Holdings), Cambridge, MA
Backend architecture for KAYAK for Business: approvals, rail distribution, payment authentication.
2023–2026
Software Engineer, KAYAK
Designed and built the approvals and trip reporting fields frameworks.
2022
Software Engineering Intern, KAYAK
Built the targeted announcements framework for the KAYAK Partner Platform. Its data model is still in production.
2021–2023
Head Graduate Teaching Assistant, Khoury College of Computer Sciences, Northeastern University
Wrote coursework and rubrics for CS 5520 Mobile Application Development and led a team of 13 teaching assistants.
2022
Software Developer Co-op, Happier (formerly Ten Percent Happier), Boston
2018–2021
Software Engineer, Fractal Analytics, Mumbai
Cuddle.AI, a natural-language business analytics product: query understanding, and multi-tenant deployment on Kubernetes.
2017–2018
Software Developer, Ingram Micro, Mumbai
IMOnline, Ingram Micro's B2B ordering platform for resellers in more than twenty countries.
Recognition
2026
Best Paper Award, EAMCON 2026 (co-author).
2026
Top writer in AI/ML and in Programming, HackerNoon, June 2026.
2025
Elevated to Senior Member, IEEE (Boston Section).
2023
Summa Cum Laude, Northeastern University.
2017
Best Student of the Year, final-year B.E. Information Technology, St. Francis Institute of Technology, Mumbai (2016–17).
2017
First prize (Shri M. M. Rai Memorial Award) for paper and poster presentation, Colloquium on “I.T. for Society”, St. Francis Institute of Technology.
Education
2023
M.S. Computer Science, Northeastern University, Boston.
2017
B.E. Information Technology, St. Francis Institute of Technology, University of Mumbai.
