| 01 MODEL ENGINEERING LLMs · RAG · NLP · CV |
02 SYSTEM ENGINEERING APIs · Data · Agents · Cloud |
03 PRODUCT ENGINEERING UX · Automation · Impact |
My goal is not to put an AI model behind a button. It is to design the complete system around the model.
Programmer Analyst @ Cognizant · Hyderabad · Dec 2024 — Present
I work on mission-critical US banking data pipelines, troubleshooting Mainframe/JCL and AB Initio ETL workflows within SLA, working with GCP BigQuery, enterprise change processes and data platforms including Teradata, Snowflake and PySpark.
Alongside production engineering, I build AI-assisted solutions that turn operational problems into repeatable workflows.
PROBLEM
↓
DATA / CONTEXT
↓
MODEL OR AGENT
↓
TOOLS / APIs
↓
VALIDATION
↓
USER-FACING RESULT
Voice-first police analytics
STT → Text Cleanup → NL → SQL → SQL Execution → Visualization
A multimodal, agent-oriented system designed for police officers to query CCTNS data using Telugu / Indian English voice input, convert natural language into SQL, execute against structured data and return visual analytics.
Core: IndicConformer · Whisper fallback · flan-T5-small · SQLova · Picard/T5 · Multi-agent orchestration
AI product prototyping under real constraints
Built and integrated a working product experience around an AI use case, including Paytm payment gateway staging integration.
Computer vision research for identity-document authentication
CNN-based forgery-detection work developed into published research, combining deep learning, feature extraction and transfer-learning approaches.
Publication: INTCEC 2024 · IEEE Xplore
Natural-language database interaction
Explores a voice-first interface for turning human questions into database operations and readable results.
| Layer | Technologies |
|---|---|
| AI | Python · PyTorch · TensorFlow · Transformers · Hugging Face · LangChain |
| GenAI | RAG · LLMs · Agents · Gemini · Ollama |
| Backend | FastAPI · Node.js · Express · REST APIs · Microservices |
| Data | SQL · MongoDB · MySQL · PostgreSQL · BigQuery · Snowflake |
| Cloud | AWS · GCP · Azure · S3 · Lambda · SageMaker · Bedrock |
| Frontend | React · Next.js · TypeScript · JavaScript · Streamlit |
| Engineering | Git · Docker · MLOps fundamentals · Production support |
┌───────────────────┐
│ HUMAN │
└─────────┬─────────┘
│ intent
▼
┌───────────────────┐
│ ORCHESTRATION │
│ agents / workflow │
└─────────┬─────────┘
│
┌───────────┼───────────┐
▼ ▼ ▼
MODEL DATA TOOLS
LLM/CV SQL/RAG APIs
│ │ │
└───────────┼───────────┘
▼
┌───────────────────┐
│ VALIDATE + GOVERN │
└─────────┬─────────┘
▼
┌───────────────────┐
│ USEFUL OUTCOME │
└───────────────────┘
Reliable AI = intelligence + context + tools + constraints + UX.
| Signal | Evidence |
|---|---|
| Production Engineering | Cognizant · banking data pipelines · SLA-driven support |
| Applied AI | CCTNS Copilot · AI-assisted production triage POC |
| Research | INTCEC 2024 · IEEE Xplore publication |
| Hackathons | AI4AndhraPolice finalist · Smart India Hackathon finalist · Intel oneAPI shortlist |
| Product Building | Paytm AI Hackathon · AP Police platforms · AI prototypes |
AI / ML ENGINEERING
↓
PRODUCTION-GRADE GENAI
↓
AGENTIC SYSTEMS
↓
SCALABLE AI PRODUCTS
↓
MEASURABLE REAL-WORLD IMPACT
I am interested in engineering environments where machine learning, software architecture and product thinking meet.
PORTFOLIO · GITHUB · LINKEDIN · EMAIL
dineshjnld.github.io/Dineshjnld · github.com/Dineshjnld · LinkedIn · dineshjnld22@gmail.com