NP
Loading Portfolio
NP/
Home01 About02 Skills03 Projects04 Research05 Education06 Contact07 ↓ Download Resume
Open to Graduate Programs & AI Roles

Naga Poojitha
Kanchukommala

AI/ML Engineer|

Building intelligent systems at the frontier of Large Language Models, Retrieval-Augmented Generation, and Agentic AI. Springer-published researcher. ECE final-year at Amrita Vishwa Vidyapeetham.

LangChainFAISSRAGLLMsAgentic AIHuggingFaceDockerSpringer 2025
Use sidebar or ↕ keys to navigate
About

The Engineer Behind
the Algorithms.

I'm an AI/ML engineer with a deep focus on building systems that think — from production-grade RAG pipelines with sub-second latency, to Agentic AI workflows that push the boundaries of what LLMs can autonomously accomplish.

My background in Electronics & Communication Engineering gives me a rare dual lens: I reason about hardware constraints, digital logic, and signal systems while engineering sophisticated AI software pipelines across the full stack.

My Springer-published research on combinatorial optimization for wireless sensor networks reflects how I approach engineering — rigorously, quantitatively, and always benchmarked against strong baselines.

Building AI that measures its impact — not just delivers it — is what separates research-grade engineering from weekend prototypes.
0%+
Hallucination Reduction
0%
Charger Count Reduction
<3%
CV Mean Absolute Error
0
Max Nodes Benchmarked
NP
Naga Poojitha Kanchukommala
AI/ML Engineer · Published Researcher
UniversityAmrita Vishwa Vidyapeetham
DegreeB.Tech. ECE · 2026
PublicationSpringer Nature · CCIS 2738 · 2025
CertificationIBM RAG & Agentic AI (8 Courses)
FocusGenerative AI · LLMs · RAG
LangChainFAISSHuggingFaceRAGAgentic AIDocker
Technical Skills

Full-Stack AI
Engineering Toolkit.

🤖
AI / ML Frameworks
LangChainLangGraphFAISSHuggingFace TransformersCrewAIAutoGenFastAPI
🧬
Machine Learning
RAG PipelinesLarge Language ModelsAgentic AIDeep LearningNLPPredictive Modeling
👁
Computer Vision
OpenCVMorphological ProcessingImage SegmentationNumPy
💻
Programming
PythonCVerilog HDLMATLAB
🔧
Developer Tools
DockerGit / GitHubJupyter NotebookVS CodeLaTeX
🔌
Hardware & EDA
Xilinx VivadoBasys 3 FPGAAnsys HFSSLTspiceAutoCAD
Research Interests
Generative AIData-Driven Model DiscoveryAI for HealthcareGraph TheorySignal DecompositionPredictive ModelingGraph Processing
Projects

Systems Built.
Impact Measured.

🧠
AI · LLM
AI-Powered RAG Chatbot
End-to-end Retrieval-Augmented Generation system with FAISS vector indexing and HuggingFace embeddings. Deployed as a containerized REST microservice.
40%+ reduction in hallucination rate vs. vanilla LLM
Sub-second query latency on custom knowledge base
GPU-independent Docker + FastAPI deployment
LangChainFAISSHuggingFaceFastAPIDockerPython
🔲
Hardware · FPGA
FPGA-Based ATM Simulation
Full ATM simulation on Basys 3 FPGA via a 12-state hierarchical FSM in Verilog HDL with 3-attempt PIN lockout and real-time 7-segment display drivers.
Deterministic latency via optimized clock-divider circuitry
Validated via behavioral simulation & post-synthesis timing
Binary-to-BCD conversion with multiplexed display
Verilog HDLBasys 3 FPGAXilinx VivadoFSM Design
🌱
Computer Vision
Morphological Seed Quality Analysis
Real-time agricultural seed grading pipeline using morphological image processing. Adaptive area-to-weight algorithm maps pixel contours to physical mass with high precision.
<3% mean absolute error across 5 seed varieties
Robust segmentation under varied lighting conditions
Configurable GUI with batch CSV export
PythonOpenCVNumPyTkinter
📡
Research · Optimization
WSN Charger Placement Optimization
Formulated and solved the Q-Coverage charger placement problem using a Blackhole Algorithm metaheuristic. Benchmarked against GA & PSO over 50–500 node topologies.
12–18% reduction in charger count vs. GA & PSO
Published — Springer Nature, CCIS Vol. 2738, 2025
Equivalent coverage with lower computational overhead
PythonBlackhole AlgorithmCombinatorial OptimizationMetaheuristics
Research

Peer-Reviewed
Publication.

✦ Springer Nature · 2025
Optimizing Charger Placement for Q-Coverage Using Blackhole Algorithm
Communications in Computer and Information Science · Vol. 2738 · Springer Nature
2025
Year Published
18%
Charger Reduction
500
Max Nodes Tested
Blackhole Algorithm
Metaheuristic inspired by black hole physics — candidates are absorbed by the best-performing solution, enabling efficient global search through combinatorial spaces.
📡
Q-Coverage Problem
Optimization challenge of placing the minimum number of wireless chargers so every sensor node receives sufficient energy — critical for perpetual WSN operation.
📊
Rigorous Benchmarking
Benchmarked against Genetic Algorithm and Particle Swarm Optimization across 50–500 node topologies, demonstrating superior efficiency and lower computational overhead.
🎯
Key Result
Achieved 12–18% reduction in required charger count vs. GA and PSO baselines while maintaining equivalent full-coverage guarantees — published in Scopus-indexed Springer CCIS.
Research Interests
Generative AILarge Language ModelsAI for HealthcareData-Driven Model DiscoveryGraph Theory & ProcessingSignal DecompositionPredictive ModelingRAG Architectures
Education & Credentials

Academic
Foundation.

B.Tech. Electronics & Communication Engineering
Amrita Vishwa Vidyapeetham, Coimbatore
Expected May 2026
Relevant Coursework
Digital Signal ProcessingVLSI DesignEmbedded SystemsDigital Image ProcessingComputer NetworksAnalog & Digital CircuitsControl SystemsProbability & ProcessesRadar Signal ProcessingSatellite CommunicationMicroprocessors & InterfacesPlanar Microwave Devices
Certifications
🏆
IBM RAG & Agentic AI Professional Certificate
IBM Skills Network · Coursera · 8 Courses
Mar 2026
🎓
Class Representative — Dept. of ECE
Amrita Vishwa Vidyapeetham · 6 Semesters
2022 – 2026
Contact

Let's Build
Something.

Open to graduate research opportunities, full-time AI engineering roles, and collaboration on cutting-edge LLM and Agentic AI projects. I bring published research-level rigor to every system I ship.

Available for Opportunities
Looking for an AI Engineer who ships research-grade systems?
From RAG pipelines that beat vanilla LLMs by 40% to Springer-published optimization research — I bring depth and delivery to every project.
Send a Message →