Principal AI Architect | Ph.D. in Computer Vision | GenAI & Multi-Agent Systems
Bridging deep academic research with enterprise-scale engineering. I architect Agentic AI systems and Deep Learning pipelines that move beyond POCs into production-grade, high-ROI innovation.
Principal Investigator for a COVID-19 thermal screening AI project, securing $3M+ in DST (Govt. of India) and VC funding.
Architected LLM-powered automation platforms for 50+ global clients (including Bosch), delivering 40% operational efficiency gains.
Founder of Codeidea.in — competitive coding & hackathon platform with 500+ students. Acquired by IVISLabs Pvt. Ltd.
LangGraph, LangSmith, OpenClaw, PyTorch, Transformers, Kubernetes, AWS.
My Ph.D. thesis, Segmentation in Compressed Document Images (University of Mysore, 2019), forms the foundation of my applied work in visual AI and OCR pipelines.
A hierarchical multi-scale network processing fundus images across four parallel branches at different resolutions, combining Dice, BCE, and centerline Dice losses with hard example mining. Achieves mean Dice of 88.72%, Sensitivity 90.78%, and AUC 98.25% across DRIVE, STARE, and CHASE_DB1.
An empirical study across five datasets quantifying the structural loss of microvascular details during downsampling.
Published in IEEE Access, vol. 12, 2024. DOI: 10.1109/ACCESS.2024.3455433
Published in Journal of Intelligent & Fuzzy Systems (JIFS), vol. 36, no. 3, pp. 2527-2544, 2019.
Developed advanced OCR and NLP models reducing manual cataloguing effort by 70%+ for e-commerce pipelines at Amazon, Flipkart, and LiveAuctioneer.
15+ years defining AI strategy, system architecture, and product roadmaps for enterprise and government clients.
Thesis: Segmentation in Compressed Document Images. Published 10+ papers in peer-reviewed journals and conferences (SCIE, SCOPUS, IEEE, IAPR).
Thesis: Indian Paper Currency Recognition Systems for Visually Impaired.
Developed core features for the EdTech platform. Technologies: Groovy on Grails, MySQL.
Strong foundation in data structures, algorithms, system design, and operating systems — the bedrock of building production-grade AI systems at scale.
Balanced BSTs, AVL rotations, segment trees, Fenwick trees. LCA, serialization, and range queries.
BFS/DFS, Dijkstra, Bellman-Ford, topological sort, strongly connected components, network flow, minimum spanning trees.
Memoization vs tabulation, knapsack variants, LCS/LIS, matrix chain multiplication, DP on trees and graphs.
Quicksort, mergesort, heapsort analysis. Binary search on answer, two-pointer, sliding window techniques.
Hash maps, collision resolution, Rabin-Karp, KMP pattern matching, trie-based solutions. Published: "Optimized string searching algorithmic technique" (2008).
Activity selection, Huffman coding, interval scheduling. Master theorem analysis, recurrence relations.
Load balancing, horizontal scaling, database sharding, caching strategies (Redis, CDN), message queues, event-driven architecture.
Encryption at rest/transit, OAuth2/JWT, rate limiting, circuit breakers. CAP theorem trade-offs, replication, failover strategies.
SQL vs NoSQL, normalization, indexing strategies, vector databases (Pinecone, Weaviate), ACID vs BASE, query optimization.
REST vs GraphQL, gRPC, API versioning, service mesh, distributed tracing, contract-first design with OpenAPI.
Feature stores, model serving (ONNX, TorchServe), A/B testing, canary deployments, MLflow pipelines, GPU-optimized inference.
Class, sequence, and deployment diagrams. GoF patterns — Factory, Observer, Strategy, Decorator. SOLID principles and clean architecture.
Process lifecycle, context switching, thread pools, green threads. Concurrency models: fork-join, actor model, CSP.
Virtual memory, paging, segmentation, TLB, cache hierarchies (L1/L2/L3). Memory-mapped I/O, NUMA-aware allocation for GPU workloads.
Mutexes, semaphores, monitors, read-write locks. Deadlock detection/prevention (Banker's algorithm). Lock-free data structures.
Raft, Paxos, gossip protocols, vector clocks. TCP/UDP internals, socket programming, WebRTC/WebSocket for real-time systems.
Docker internals (cgroups, namespaces), Kubernetes scheduling, pod lifecycle, resource quotas, HPA for ML inference workloads.
Disk scheduling, journaling, RAID levels, I/O multiplexing (epoll, kqueue). Object storage (S3) for ML artifact management.
Interactive proof-of-concepts showcasing applied AI capabilities and engineering depth.
Autonomous research agents that retrieve, synthesize, and critique domain-specific documents in real-time.
LLM-powered orchestration for enterprise document processing with human-in-the-loop validation.
Interactive retinal vessel segmentation with hierarchical multi-scale attention visualization.
Autonomous agents for third-party risk assessment, compliance monitoring, and report generation.
End-to-end OCR and NLP pipeline for automated content extraction from unstructured documents.
Custom evaluation harness for benchmarking RAG quality, hallucination rates, and latency across models.
Real-time messaging, voice, and video platform optimized for low-latency communication across devices.
Algorithm visualizations, real-time problem-solving environments, and university hackathons for 500+ students.