Hi, my name is

Rishi

AI Engineer · Agentic AI · Full Stack Software Engineer

I build production-grade AI systems and high-throughput backend platforms—specializing in Agentic AI (tool-using agents, orchestration, evaluation), RAG pipelines, and cloud-native engineering. I’ve shipped multiple AI MVPs, scaled data/embedding pipelines to tens of millions of records, and delivered enterprise workflows with strong reliability, security, and observability.

About Me

I’m an AI Engineer and Full Stack Software Engineer focused on building end-to-end AI products—from data ingestion and retrieval to agent orchestration and UI delivery. My recent work spans Agentic AI systems for EDI automation (hybrid rules + LLMs), multi-agent workflows, and RAG-based knowledge systems that turn operational documentation into fast, accurate answers.

I’m hands-on across the stack: Python/FastAPI, Spring Boot, React/Angular, Kafka, Kubernetes, Terraform, and cloud AI platforms (Vertex AI, AWS, Azure). I emphasize predictable agent behavior (stateful graphs), measurable quality (evals/scorecards), and operational readiness (tracing, cost controls, and guardrails).

Experience

AI Engineer

Vision Benefits of America (VBA)
Milwaukee, USA
  • Engineered AI-assisted EDI mapping workflows using LLMs (LlamaIndex) to automate companion doc parsing, field mapping, and schema alignment.
  • Built validation agents that generate mapping scorecards (row validity, length/type mismatches, accumulator checks), reducing manual QA cycles.
  • Prototyped agent-driven import assistants for enrollment files, automating code mapping, relationship/tier setup, and dependent validation.
  • Implemented RAG knowledge retrieval from setup guides, mapping specs, and conversion docs to accelerate onboarding and reduce reliance on templates.
  • Improved traceability and tuning with Langfuse across multi-agent workflows (token usage, latency outliers, drift signals, prompt quality).
  • Developed deterministic multi-step reasoning pipelines using LangGraph state graphs for predictability, debuggability, and error recovery.

Founding Engineer (Freelance)

Onward Platforms
Remote
  • Designed and implemented MCP-based AI agents enabling a social “@agent” invocation experience for real-time, tool-augmented responses.
  • Built orchestrated agent workflows with Python/FastAPI and cloud AI services, integrating LLMs, custom tools, and retrieval for higher relevance.
  • Developed reusable agent components: memory, planning, tool routing, and structured outputs for plug-and-play agent development.
  • Integrated enterprise connectors and APIs (e.g., Google Workspace, ServiceNow) to support automation use-cases with permissions-aware execution.
  • Implemented production considerations: rate limiting, failure handling, logging/tracing, and prompt/version management for safe iteration.

Applied GenAI Engineer (Freelance)

100x Engineers
Remote
  • Developed a RAG pipeline for a chatbot using LangChain and cloud data tooling to deliver product insights with strong retrieval precision.
  • Built scalable APIs using Python/FastAPI; implemented keyword generation, delta-checks for data consistency, and robust error handling.
  • Improved pipeline performance and reliability with testing (Pytest) and data processing optimizations (Pandas/NLTK).

Software Engineer 2

American Tire Distributors
USA
  • Led a production RAG pipeline: embedded queries, retrieved context via vector search (cosine similarity), and generated grounded responses with Gemini.
  • Built embeddings pipeline for a 20M+ record dataset using batching + Dataflow into BigQuery, enabling fast semantic retrieval at scale.
  • Modernized cloud AI infrastructure: Kubernetes + Terraform deployments, rate limiting for LLM endpoints, and reproducible multi-environment delivery.
  • Designed a high-throughput EDI processing system handling 100,000+ daily transactions; improved performance and reduced error rates through modular design.
  • Automated scheduling jobs for 50+ partners using Python and Go, reducing operational overhead and improving reliability.
  • Delivered web experiences with React/TypeScript and integrated backend validation + Kafka consumers for real-time anomaly reporting and lineage tracking.
  • Optimized cluster sizing and scaling to handle peak fanout traffic while improving cost efficiency.

Software Engineer

Digital Bank of Singapore
India
  • Built websites, micro-frontends, and landing pages end-to-end, supporting high-traffic, customer-facing experiences.
  • Created a reusable micro-frontend component library used across public sites, improving development velocity and consistency.
  • Improved scalability and reliability through cloud deployments and performance-focused engineering practices.
  • Ensured strong cross-browser compatibility and stable user experience across major browsers.

Associate Software Developer

Syneffo Solutions
India
  • Improved search performance using Elasticsearch indexing strategies to reduce query time and enable faster filtering/sorting.
  • Processed large volumes of unstructured data using cloud data tooling to support analytics and downstream systems.

Skills

Agentic AI

  • Agent Orchestration: LangGraph (state graphs), LlamaIndex Workflows, LangChain
  • Tool Use: function calling, tool routing, structured outputs (JSON Schema / Pydantic)
  • RAG: chunking, hybrid retrieval, reranking patterns, grounding, citations
  • Multi-Agent Patterns: routing, fan-out/fan-in, reflection loops, handoffs, debate
  • MCP (Model Context Protocol) & tool registries
  • Evaluation: scorecards, regression checks, quality gates, LLM-as-a-judge patterns
  • Safety & Security: PII-aware workflows, redaction/anonymization, permissions-aware tools

Observability & Reliability

  • Tracing/Telemetry: Langfuse, OpenTelemetry concepts, prompt/version tracking
  • Monitoring: Prometheus, Grafana, CloudWatch
  • Operational Guardrails: rate limiting, retries, DLQs, timeouts, circuit-break patterns
  • Cost Controls: token budgeting, caching, deterministic-first hybrid architectures

AI / ML

  • LLMs: GPT, Gemini, LLaMA, Mistral, DeepSeek
  • NLP, Prompt Engineering, Transformers
  • Model Integration & Deployment
  • PyTorch, TensorFlow (working knowledge)

Cloud AI / MLOps

  • GCP: Vertex AI, BigQuery, Dataflow, Cloud Functions, GKE
  • AWS: EC2, EKS, S3, Lambda, Bedrock
  • Azure: Azure AI / ML, App Services
  • CI/CD: Jenkins, GoCD, GitHub Actions
  • Terraform, Docker, Kubernetes

Backend / Data

  • Python (FastAPI, Flask, Django), Go, Java (Spring Boot)
  • Event-Driven Systems: Kafka, RabbitMQ
  • Databases: PostgreSQL, MySQL, MongoDB
  • Search: Elasticsearch
  • Data Pipelines: Airflow, PySpark (working knowledge)

Frontend

  • React, Next.js, Angular
  • TypeScript / JavaScript, HTML5, CSS3, SASS
  • UI: Material UI, Tailwind/Bootstrap (as needed)
  • Testing: Playwright

Tools

  • Git (GitHub/GitLab/Bitbucket), Jira
  • Swagger/OpenAPI, Postman
  • Maven, Gradle, npm
  • VS Code, IntelliJ, Eclipse
  • Auth: OAuth/JWT/Okta

Education

Master of Science, Computer Science

University of North Carolina at Charlotte, NC, USA
GPA: 4.0

B.E. in Computer Science

Vardhaman College of Engineering, Hyderabad, India

Projects & Certifications

Agentic EDI Mapper (Hybrid AI + Rules)

  • Built agentic workflows to parse companion specs, propose mappings, validate schema constraints, and generate structured outputs.
  • Implemented deterministic-first matching with LLM assistance for ambiguous cases to improve reliability and reduce cost.
  • Added evaluation scorecards and traceability (Langfuse) to monitor quality, latency, and drift.

RAG Chatbot with Vector Search

  • Designed a RAG pipeline using vector search for fast semantic retrieval and LLM prompting for grounded, context-aware answers.
  • Implemented embedding generation workflows and optimized ingestion for large-scale datasets.
  • Added guardrails: rate limiting, caching, and robust failure handling for production stability.

EDI Integration – Event-Driven Data Exchange Platform

  • Developed a high-throughput EDI integration service using Spring Boot and Kafka producers/consumers for scalable message processing.
  • Designed fault-tolerant pipelines (retries, dead-letter queues, fallbacks) for reliable processing under dependency failures.
  • Improved throughput with multithreading and async execution patterns while maintaining traceability.

AI-Driven Validation & Production Readiness Checks

  • Built validators to simulate downstream outcomes, flag anomalies, and recommend remediation prior to scheduled execution.
  • Combined rules-based checks with LLM reasoning to handle edge cases and reduce human review cycles.

Frontend Platforms & Micro-Frontends

  • Built reusable component libraries and micro-frontend experiences to accelerate feature delivery and maintain consistency.
  • Delivered performance-focused UX with stable cross-browser behavior and production deployment workflows.

Certifications & Highlights

  • Google Professional Machine Learning Engineer
  • Machine Learning Specialization — deeplearning.ai
  • AWS Certified Solutions Architect — Professional
  • Kubernetes Certified Administrator (CKA)
  • Hackathon finalist (Top 5) — DBS Hackathon
  • Megathon winner — IIIT Hyderabad / IBM challenge
  • Gladiator Recognition Award — DBS Bank
  • BEC Certification (Vantage Level) — Cambridge English

Contact

Get In Touch

I’m open to AI Engineering, Agentic AI, and Full Stack opportunities—especially roles that involve building reliable AI systems with strong retrieval, orchestration, and observability. If you’d like to collaborate or discuss an opportunity, please reach out.