I’m a principal-level engineer with 14+ years across distributed systems, cloud platforms, automation, CI/CD, observability, production reliability, and customer-facing delivery.
Career direction: Moving into Forward-Deployed AI, AI Platform Engineering, and LLMOps / AI Reliability.
Now I extend that foundation through independent, inspectable AI systems spanning grounded RAG, multi-model and agentic workflows, LLM evaluation, reliability, security, cost controls, observability, and release governance.
Most recently: Oracle Principal Member of Technical Staff · Technical leadership: Led and mentored 11+ engineers
Explore selected AI systems · LinkedIn · Email
After my Oracle role concluded in April 2026, I chose a focused period for:
- Full-time portfolio engineering across applied AI systems
- Evidence-led AI platform development with inspectable architecture, evaluation, reliability, and release controls
- Targeted professional upskilling in Forward-Deployed AI, AI Platform Engineering, LLMOps / AI Reliability, and end-to-end MLOps
This focused period also includes a personal break; these systems are independent work, not employer projects.
From customer ambiguity to production learning—with trust and runtime discipline at every stage.
Read the delivery model as text
- Understand: Discover customer problems and frame requirements, constraints, risks, data, stakeholders, and outcomes → decision brief
- Design & Deliver: Define boundaries, services, APIs, data and trust decisions; build workflows, integrations and platform capabilities → system boundary
- Assure & Ship: Run system tests and LLM evaluations for quality, security and performance; release through CI/CD, gates, canaries and rollback → release evidence
- Run & Improve: Operate with observability, SLOs, incidents, reliability and cost controls; when systems are deployed, learn from production feedback, adoption, and improvement → learning backlog
- Across every stage: Trust, provenance, evaluation, security, release governance and runtime discipline
- Feedback loop: Learning informs the next discovery cycle
01 / CiteVyn
Answers a practical question: Can I trust this AI-generated answer—and can I trace every claim to an authoritative source?
What it delivers: Citation-grounded answers with refusal controls and evaluation-gated index promotion.
Grounded RAG · Citations · Strict refusal · Versioned retrieval ·
Evaluation gates
State: Public portfolio with production-oriented controls; current deployment, provider, index, users, and adoption are not asserted.
02 / Quorum-AI
Answers a practical question: How can I compare several models while controlling cost, provenance, disagreement, and degraded execution?
What it delivers: Parallel analysis, critique, and synthesis with pre-run cost approval and provider/simulation disclosure.
Multi-model orchestration · Critique and synthesis · Cost gates ·
Provider provenance · Degraded modes
State: Public portfolio; execution may use live providers, failed-slot fallbacks, or simulation; providers and adoption are not asserted.
03 / SaafSaans
Answers a practical question: What does current air quality mean for my plans—and when might conditions improve?
What it delivers: Persona-scoped guidance with citations, feed provenance, and labelled live, deterministic, or sample modes.
Grounded guidance · Citations · Feed provenance · Labelled fallbacks ·
Injection guard
State: Public, non-clinical guidance; heuristic risk scoring has documented feed, pollutant-coverage, and WAQI-use limitations.
04 / NarraTwin AI
Answers a practical question: How can project knowledge become an audience-specific walkthrough without inventing unsupported claims?
What it delivers: Grounded scripts with citations, claim evaluation, consent checks, and pre-generation release gates.
Grounded scripts · Citations · Claim evaluation · Consent ·
Release governance
State: Local/mock Phase 1 under No-Go; single-node restart recovery; no hosted deployment, provider-backed media generation, real video, or public distribution.
Across Oracle, Amazon, LimeRoad, Mobileum, Snapdeal, and Subex, my employer-backed experience spans distributed systems, cloud analytics, data platforms and microservices; automation architecture, CI/CD and release governance; observability, canaries, runtime validation and production reliability; and customer-facing delivery, integration, UAT and stakeholder collaboration across devices, e-commerce, search, payments and telecom systems.
Most recently, I served as an Oracle Principal Member of Technical Staff from April 2019 to April 2026, with OCI as my strongest professional cloud.
Selected outcomes from my Oracle tenure, separate from my independent portfolio:
- MTTD reduced approximately 35% through telemetry and dashboards.
- Targeted release workflows reduced cycle time approximately 25%.
Trust before deployment. Validate production reliability continuously. Shift left by default; shift right by design.
Evaluation, observability, degraded modes, security, and release gates are design concerns.
Open to senior/principal opportunities in Forward-Deployed AI, AI Platform Engineering, and LLMOps / AI Reliability.
Also open to closely aligned roles across AI Quality & Test Engineering, AI/ML SDET and Test Architecture, and AI Platform & DevOps.
Based in Bengaluru, India; open to global relocation and international travel.