About .
I'm an AI & Backend Engineer and a Forward Deployed Engineer at CDX Global, supporting AI-assisted cybersecurity workflows and agent access controls for enterprise clients. I hold an M.S. in Computer Science from the University of the Pacific (May 2026).
My journey started in Engineering at BITS Pilani, but my passion for data and intelligence led me to code. I've since built StoryAware, a hierarchical RAG platform, self-critiquing agents like StageLoop, and NestFit, a multimodal product that checks furniture fit at true scale.
Beyond Code
When I'm not building agents, I'm likely watching Real Madrid dominate Europe (Hala Madrid! ⚽️🇪🇸), playing badminton, or analyzing chess positions.
Education
M.S. in Computer Science
University of the Pacific
May 2026 • Stockton, CA
GPA: 4.0
B.E. in Engineering
BITS Pilani (Hyderabad Campus)
May 2019 • Hyderabad, India
Experience
Forward Deployed Engineer
Client: IBM · AI-Assisted Cybersecurity Workflows
Feb 2026 — Present- Supported AI-assisted threat intelligence workflows that bring together indicators of compromise, public social and professional profile information, and approved dark-web intelligence to help security teams understand employee exposure and build contextual threat profiles.
- Worked with customer and security teams on external attack-surface assessment and authorized social-engineering scenarios, mapping how help-desk impersonation and excessive privileges could expose critical identity systems.
- Supported AI-driven phishing and vishing simulation use cases, including voice impersonation, and connected assessment findings with employee awareness training and remediation priorities.
- Helped define which data and tools each AI agent could access through identity and access management (IAM) and least-privilege controls; planned containment workflows to revoke access, lock affected accounts, and stop unsafe automation.
Client: Sky Transport Solutions · Fleet Compliance Platform
Nov 2025 — Feb 2026- Developed Python and FastAPI APIs with PostgreSQL to automate fleet compliance workflows, reducing manual data entry by 25%.
- Built an OCR and LLM pipeline to extract data from driver records and IFTA receipts, improving extraction accuracy by 15% and shortening document review time.
- Connected compliance systems through REST APIs, optimized SQL queries, and introduced asynchronous request handling; built a LangChain and pgvector RAG prototype that shortened regulation lookup time.
- Worked in a five-engineer team to containerize services with Docker, automate unit tests through GitHub Actions, and document APIs with OpenAPI / Swagger, reducing deployment setup time by 25%.
Lead Researcher (AI/ML Engineering Focus)
- Designed StoryAware, a domain-aware book intelligence platform with hierarchical RAG and intent-based routing, improving grounding quality on dense philosophical texts by ~30%.
- Built a production GraphRAG pipeline with ontology-constrained extraction, hybrid vector and keyword search, and citation-first answers using FastAPI, PostgreSQL, and pgvector, reducing retrieval latency by ~35%.
- Supported ~50 weekly active users who reported study time savings of 5–10 hours per week; the platform uses 8B and 32B models and is used across ISKCON temples in India.
- Developed live webcam emotion recognition with multi-class classification, improving inference consistency by ~13%, integrated into a browser-based TypeScript prototype.
Software Engineer (ML & AI)
- Built ML and NLP pipelines with scikit-learn, XGBoost, spaCy, and Hugging Face models; applied feature engineering, class weighting, and imputation to datasets with up to 40% missing fields.
- Productionized models through FastAPI, Docker, MLflow, and CI/CD; collaborated on SQL and PySpark pipelines for ingestion and transformation.
- Developed LangChain RAG pipelines using Pinecone, ChromaDB, BM25, and reranking; introduced Redis semantic caching that reduced LLM API costs by 35% and cut cached response times from ~3 seconds to under 100 ms.
Software Developer
- Built Java Spring Boot microservices and REST APIs for onboarding, learning, and quiz workflows; optimized MySQL, MongoDB, and Redis to reduce API latency by 20%.
- Used AWS EC2, S3, SQS, and CloudFront for hosting and content delivery; improved Jenkins and Docker CI/CD pipelines, reducing deployment time by 35%.
Projects
Everything I've built — tap a project for the full case study or live link.
StoryAware
Grounded, hierarchical, graph-aware book intelligence platform with multi-level retrieval and source-first answers.
Emotion Recognition
Real-time emotion detection with computer vision models.
Robo Wars Combat Robot
3rd place hardware team build focused on durability and performance.
SCARA Cleaning & Scrubbing Robot
Two-arm cleaning robot with a powered scrubber, vacuum routing, and vertical head movement for stove-surface cleaning; sizing drives, reductions, and arm packaging.