01
Blog Hub
A blog management and publishing platform for readers, authors and administrators.
BLOGGING & PUBLISHING PLATFORM · 2026
Open project →
GENERATIVE AI & AGENTIC AI ENGINEER
I build AI agents and RAG systemsthat work inside real products.
Generative AI and Agentic AI engineer working across LLMs, retrieval-augmented generation and tool-calling agent workflows — then shipping them full-stack, from data model to service layer to interface.
AGENT & RAG PIPELINE LIVE LOOP
- BASED IN
- Hyderabad, India
- FOCUS
- RAG · agent workflows · evaluation
01
About & how I work
I am a computer science engineer from Hyderabad who ended up in applied AI because the hard parts turned out to be ordinary engineering parts: deciding what the model is allowed to see, what it is allowed to do, and how anyone can tell whether it worked.
Most of my work sits in two places. Retrieval — building the index, the hybrid search over it, and the prompt that answers only from what came back. And agents — giving a model typed tools, a bounded loop, and a definition of done that code can check.
I finish my B.Tech in Computer Science at Mohan Babu University in 2026 with a CGPA of 9.2, alongside internships at TMaxEdTech and ExcelR Edutech and two platforms shipped end to end. Underneath the AI work there is a habit from 200+ DSA problems in Java: get the data structure right before you get clever.
- 01
Decide what the model sees
Instructions, retrieved passages, tool schemas, history — and what gets cut. Most answer-quality problems are context problems, not model problems.
- 02
Make the output a contract
A fixed JSON schema at the boundary, so downstream code depends on a shape instead of parsing prose that changes with the weather.
- 03
Keep the loop bounded
Typed tools, a step ceiling and fallback handling, so a run that goes wrong fails visibly instead of spending quietly.
- 04
Score it, then change one thing
A generator–critic pass and a scored set beat a vibe check. Change one variable at a time so an improvement can be shown rather than claimed.
- 05
Move slow work off the request path
Notifications, scheduled pulls and database writes belong in n8n flows and background jobs, not in the user's wait.
02
What I build
Deciding what goes into the model's window — instructions, retrieved passages, tool schemas, conversation history — and what gets trimmed when it will not all fit. This is where most quality lives.
Retrieval-augmented generation over vector and relational stores, combining keyword and semantic search so exact terms — IDs, names, skills — and meaning both get found, then merging the ranked lists.
Tool calling, ReAct and plan-execute loops, and multi-agent orchestration with LangGraph and n8n — with typed tools and a step budget so the loop terminates on purpose.
Eval harnesses, schema validation, generator–critic filtering, model rotation, throttling and fallback handling — the parts that decide whether a feature survives its first real week.
03
Technical stack
Every item is listed with what it is and where I use it — no clicking required.
DEPTH BY AREA
Retrieval & RAG
hybrid search, grounded answersAgentic workflows
tool calling, ReAct, LangGraphFull-stack delivery
React, Node.js, SQLAutomation (n8n)
scheduled & notification flowsModel serving
vLLM, Ollama, self-hostedDSA in Java
200+ problems on LeetCode04
Selected work
Open a project to read the case study — problem, approach, architecture and outcome.
02
HireGen AI
An AI recruitment platform that runs the whole hiring lifecycle, from job posting to scored candidate evaluation.
RECRUITMENT & ASSESSMENT · 2026
Open project →
03
Vivāha Atelier
A marketing site for a luxury wedding invitation brand, built to turn browsing into enquiries.
MARKETING PLATFORM · 2026
Open project →
05
Experience
FEB 2026 — JUL 2026
Generative AI & Agentic AI Intern TMaxEdTech
- Built interactive interfaces to visualise agent workflows, prompt runs and evaluation results.
- Developed REST APIs and service logic in Python to orchestrate LLM agents, tool calling and RAG pipelines.
- Designed and integrated schemas to persist conversation history, agent state and retrieval embeddings.
MAY 2025 — JUN 2025
Java Programming Intern ExcelR Edutech (APSCHE)
- Completed a six-week internship focused on Core Java and object-oriented design.
- Built mini-projects and coding exercises that turned into a steady DSA practice habit.
06
Education
01B.Tech, Computer Science & Engineering Mohan Babu University, Tirupati
02Higher Secondary, Class XII Sri Chaitanya Junior College
03SSC, Class X Sri Chaitanya E-Techno School
07
Certifications & achievements
01Internship
Generative AI and Agentic AI
TMaxEdTech · Feb–Jul 2026
02Certification
Java Foundation Certification
Infosys Springboard
03Internship
Java Programming Internship Certificate
ExcelR Edutech · APSCHE
04Achievement
Solved 200+ data structures & algorithms problems
LeetCode · Java
05Achievement
National and district-level basketball player
School & college tournaments
08
My journey
From first lines of Java to shipping AI products.
2022 — 2026
B.Tech, Computer Science & Engineering
Started at Mohan Babu University, Tirupati. Java and DSA came first — 200+ LeetCode problems — and the habit of naming the data structure before writing the loop stuck around.
MAY — JUN 2025
Core Java, properly: ExcelR Edutech
A six-week Java internship under APSCHE, spent on OOP and small projects. The first code I wrote to somebody else's specification instead of my own.
FEB — JUL 2026
Agents in production: TMaxEdTech
Python REST APIs orchestrating LLM agents, tool calling and RAG pipelines, the schemas holding conversation history and embeddings, and interfaces for watching prompt runs and evaluation results.
2026
Two platforms, two problem shapes
HireGen AI — a generator–critic assessment engine with a real service layer around it. Vivāha Atelier — a catalogue and enquiry pipeline where the AI is not the point and the polish is.
NOW
Building, measuring, job-hunting
Looking for an entry-level Generative AI or Agentic AI engineering role, and continuing to build retrieval and agent projects in the meantime.
09
Why hire me
Four things that show up across every project.
AI inside real products
Every model I have shipped sits in a complete product with auth, a database and an admin path — not an isolated notebook demo.
Full-stack ownership
Front end, API, schema and the AI layer. I built the data model and the interface for both platforms myself.
Reliability by design
Model rotation, throttling, fallback handling and schema validation ship with the feature, not after the first outage.
Fundamentals underneath
200+ DSA problems in Java, plus DBMS, OOP and computer networks. The AI layer is new; the engineering is not.
10
Lab
Smaller experiments and explorations that don't need a full case study.
Agentic AI
Tool-calling ReAct loops
Small agent prototypes exploring plan-execute loops and multi-agent orchestration with LangGraph and n8n.
Prompt engineering
Structured-output reliability
Testing context engineering, guardrails and schema-forced output across GPT-4o, Claude and Gemini.
Retrieval
Hybrid retrieval playground
Comparing hybrid retrieval strategies over VectorDB and Postgres for grounding answers in real documents.
Local LLMs
Self-hosted serving
Running Llama 3, Mistral and Qwen through vLLM and Ollama to compare latency and cost trade-offs.
Evaluation
Prompt regression sets
Small scored test suites that run on every prompt change, so a regression shows up before a user finds it.
Automation
n8n workflow patterns
Reusable scheduled-pull and notification-routing patterns lifted out of BizIntel and Royal Occasions.
11
Curious about my work? Just ask.
A keyword-matched assistant grounded only in this page's real content — instant answers about the projects, stack and experience, with no live model involved.
Portfolio Assistant
Hi — I'm a local, keyword-matched assistant built only from the real content on this page. Ask me about the projects, the stack, the internships, or how to get in touch.
Try asking
AVAILABLE FOR HIRE
Let's build somethingthat holds up.
Open to entry-level Generative AI and Agentic AI engineering roles, in Hyderabad or remote.
CONTACT DETAILS
PHONE
GITHUB