Danish Ahmed

AI Engineer

I design and build scalable, secure end-to-end systems that turn data and AI into measurable business outcomes. My work focuses on architecting reliable, production-grade solutions that solve complex, real-world problems.

Experience

AI Engineer· DerivexpandJul 2026 - Present

Building AI and automation systems that eliminate manual operational bottlenecks and improve engineering reliability.

Approach

Built an end-to-end agentic pipeline that handles data access requests autonomously — gathering and consolidating requestor data, formatting output to organizational standards, and redacting third-party PII before routing the completed package for compliance review and sign-off. Separately, authored full technical documentation for an existing production system that had none; became the primary maintainer — triaging inbound issues from stakeholders, resolving defects directly, and escalating to the wider team where needed.

Impact

Data access request turnaround reduced from 7 days to under 15 minutes. Production system gained a documented, maintained baseline — stakeholder issues now routed and resolved systematically rather than ad hoc.

PythonAgentic AILLMsNLPMLOpsGCPDockerKubernetes
Data & Automation Engineer· OMNI GroupexpandApr 2026 - Jun 2026

First analyst hired; built the company's entire data and reporting function from scratch on a 150+ table SQL database with millions of rows.

Approach

Designed an executive arrears dashboard covering 3,000+ live cases with drill-down into arrears, catch-up balances, and assignment-stage pipeline. Built an API layer to serve query results to reporting tools, and engineered custom SQL logic with senior leadership to surface and reconcile cases not accounted for by existing systems.

Impact

Exposed £60K in previously untracked arrears and catch-up payments, tracked ~£500K in customer payments; dashboard adopted operationally by senior leadership and two business teams.

SQL ServerNode.jsPower BIDAXExcel
Software Engineer· KingScote InformaticsexpandFeb 2025 - Mar 2026

Built and shipped AI-native backend systems and agentic pipelines spanning data infrastructure, production incident management, and enterprise AI tooling.

Approach

Designed and built AI-native backend systems and agentic pipelines across multiple domains: a unified data access layer serving multiple sources through a single secured REST API; an automated incident analysis pipeline with timeline reconstruction, root cause analysis, and post-mortem generation; an agentic multi-platform content publishing system with human-in-the-loop approval; a conversational AI booking agent handling natural-language requests end-to-end; a multi-agent customer support triage system with RAG-powered responses; and a permission-aware enterprise knowledge assistant with role-based document retrieval and cited, auditable answers.

Impact

Unified data layer adopted across all operational teams, providing consistent access to both SQL and NoSQL sources through a single interface. Incident analysis turnaround reduced from 1 day of manual work to under 10 minutes, fully automated. Booking agent served ~3,000 customers weekly with 24/7 availability. Authorization and compliance constraints enforced at the infrastructure level — sensitive data filtered before reaching AI models.

PythonFastAPILangChainLangGraphLLMsRAGPineconePostgreSQLRedisDocker
Data Scientist· KingScote InformaticsexpandNov 2024 - Jan 2025

Applied end-to-end ML pipelines to energy forecasting and clinical survival prediction, with emphasis on model interpretability and production-ready model evaluation.

Approach

Built a time-series forecasting pipeline systematically comparing ARIMA, Prophet, and a stacked LSTM on 35K hourly energy demand records, with automated rolling-window backtesting and an interactive dashboard for model comparison and residual analysis. Separately, developed a clinical survival prediction model for post-HCT patients — benchmarked six ensemble and gradient boosting algorithms, optimized the best performer via Optuna with recall prioritized to minimize false negatives, and applied SHAP values and decision tree visualization for clinical interpretability. Conducted unsupervised patient clustering to surface sub-groups within the high-dimensional clinical dataset.

Impact

LSTM achieved 1% MAPE vs. ARIMA baseline — 83% forecasting error reduction. Post-HCT model achieved 0.94 recall and 0.80 precision after tuning; clustering surfaced clinically meaningful patient sub-groups consistent across multiple algorithms.

PythonLightGBMXGBoostScikit-learnPyTorchSHAPLSTMARIMAPandas
Teaching Assistant· Delta Sigma TechnologiesexpandJul 2022 - Aug 2022

Mentored 120 students in Python and applied ML across guided lab sessions.

Approach

Ran labs on preprocessing, visualization, and ML algorithms; covered engineering-level mathematics in Python — calculus, linear algebra, and signal processing. Reviewed code and gave targeted feedback to improve correctness.

Impact

Students independently completed data analysis, ML, and applied mathematics assignments by end of cohort.

PythonPandasNumPyMatplotlibSeabornScikit-learnSciPySymPy

Projects

MAST Metacognitive Monitoring for Agentic Software Developmentexpand2026

Non-intrusive observer layer for LangGraph multi-agent systems that detects 7 behavioral failure signatures and auto-recovers via checkpoint rollback — 100% task completion across 24 benchmark runs.

Approach

ObserverEngine evaluates 7 MAST failure signatures (stagnation, tool loops, context overload, role violations, etc.) over a sliding event window; when a signature persists above threshold, injects a natural-language recovery hint and restores the LangGraph MemorySaver checkpoint to the last stable state. Attaches asynchronously without modifying agent prompts, roles, or tool definitions.

Impact

100% task completion across 12 matched observer/baseline pairs; event-count reduction in 7/12 cases (median −14 events); up to 69% token savings on longer runs.

PythonLangGraphLangChainOpenAIMulti-agent SystemsObserver Pattern
Table Lens — Natural Language Business Intelligenceexpand2026

Full-stack BI tool that lets users query databases in plain English — multi-agent pipeline handles schema discovery, SQL generation, result interpretation, and chart creation end-to-end.

Approach

Multi-agent architecture with specialized agents per task: schema discovery over enterprise-scale databases, SQL generation and validation, plain-English result summarization, and chart generation from natural-language descriptions. Read-only access with rate-limiting; Next.js frontend with ECharts for saved dashboards; Supabase (Postgres + pgvector) for storage.

Impact

Demo runs against a production-grade property and casualty insurance database — 50 tables, 2,215 columns, 5 million rows — deliberately chosen to stress-test inconsistent naming and denormalized schemas. 80+ automated tests.

PythonFastAPILangChainNext.jsSupabasePostgreSQLpgvectorEChartsOpenRouter
Nizam — AI Operating Systemexpand2026

Private AI operating system on a live VPS — 10 autonomous agents each own a domain of personal life or business, controlled entirely through Discord.

Approach

10 agents, 3 layers: personal layer (sysadmin, knowledge curator, personal assistant), a dedicated researcher agent for multi-source due diligence, an investor agent with IBKR read access and compliance-aligned screening, and a business C-suite (Chief of Staff, CFO, COO, CTO, CMO). Each calls domain-specific MCP servers backed by PostgreSQL + pgvector; all LLM calls route through LiteLLM + Redis to OpenRouter — models are plug-and-play. Knowledge vault ingests from multiple sources with approval-gated writes. Full observability via Prometheus, Grafana, Loki, and LangFuse.

Impact

One VPS, one Discord server — complete personal and business operating layer covering infrastructure, knowledge curation, finance, habits, tasks, research, investment due diligence, and a full business C-suite.

PythonMCPHermesLiteLLMPostgreSQLpgvectorRedisDiscordPrometheusGrafanaLokiOpenRouter
NL for NoSQL — Natural Language Query Agentexpand2026

Agentic chatbot that lets users query a MongoDB Atlas database in plain English, combining vector search, full-text search, and document retrieval in a stateful ReAct loop.

Approach

LangGraph StateGraph orchestrates an LLM agent over three tools — semantic plot vector search, Atlas full-text title search, and full document retrieval — with MongoDBSaver checkpointing full conversation state and MongoDBStore persisting long-term user memory across sessions. Dual interface: Streamlit web UI and terminal CLI.

Impact

Fully stateful multi-turn querying over the sample_mflix dataset; agent autonomously selects the right retrieval strategy per question without user knowledge of query syntax.

PythonLangGraphLangChainMongoDB AtlasVector SearchOpenAIStreamlit
Resumate — AI Job Application Systemexpand2025

CLI job application tracker with AI-powered ATS scoring, tailored resume and cover letter generation, and full Google Cloud sync.

Approach

Typer and Rich CLI with Pydantic-validated schemas across five modules: application tracking with status pipeline and stats dashboard, ATS fit scoring against job descriptions, tailored resume and cover letter generation across multiple tones, skill gap analysis with learning path generation, and Google Sheets/Docs/Drive sync via OAuth2.

Impact

End-to-end pipeline from tracking applications to generating tailored documents and syncing them to cloud storage — covering the full job search workflow in one tool.

PythonTyperOpenAIPydanticGoogle Cloud APIsRich
12 more projects
SignWave — Real-Time Sign Language Recognitionexpand2024

Wearable sensor glove that translates ASL hand gestures into speech in real time — 98% classification accuracy with sub-second end-to-end latency.

Approach

Flex and IMU sensors embedded in a glove capture finger bend angles and wrist orientation on an ESP32, streaming raw motion data over WebSockets to a cloud-hosted Random Forest classifier trained on labelled gesture sequences. Classified output is passed to a text-to-speech engine and played back in real time, closing the loop from gesture to spoken word without any manual trigger.

Impact

98% classification accuracy across the ASL gesture set with sub-second latency from gesture to speech; end-to-end pipeline runs on commodity hardware with no GPU.

ESP32PythonRandom ForestWebSocketsIoTScikit-learnText-to-Speech
Last-Mile Delivery Optimizationexpand2025

SQL-driven diagnostic study on a large last-mile delivery dataset isolating peak-hour delay patterns, with a clustering-based batching simulation to estimate dispatch efficiency gains.

Approach

PostgreSQL CTEs, window functions, and grouped diagnostics decompose delays by hour, traffic density, weather, and city; dataset split into a star schema (rider, city, vehicle, route dimensions + delivery fact table). K-Means clustering over geospatial order points with time-window constraints simulates batch dispatch. Power BI report for executive storytelling.

Impact

Peak-hour bottleneck patterns fully evidenced across operational dimensions; batching simulation produced directional efficiency gain estimates under available data constraints.

PostgreSQLPythonKMeansPower BISQLGeospatial Analysis
AI-Based Network Intrusion Detection System — Architecture Designexpand2026

Architecture-centric design for a scalable AI NIDS that co-designs software design patterns with distributed microservices for production reliability.

Approach

Combined creational, structural, and behavioral software design patterns (factory, decorator, strategy, observer) with microservices, event-driven messaging, and pipe-and-filter decomposition inside the detection path.

Impact

Implementation-ready architecture addressing four gaps in classical AI-NIDS — scalability, maintainability, extensibility, and adversarial robustness — with evaluation framework against NSL-KDD and CICIDS-class datasets.

Software ArchitectureDesign PatternsDeep LearningMicroservicesCybersecurityPython
Vision-based Autonomous Vehicleexpand2025

Computer-vision robot with real-time obstacle detection and point-to-point navigation.

Approach

YOLOv5 on Jetson Nano fused with ultrasonic and IR sensors via Arduino and ESP32 for dynamic speed control.

Impact

Accurate point-to-point navigation in cluttered environments at real-time inference speed.

YOLOv5Jetson NanoArduinoESP32Computer Vision
Blockchain-based Voting Systemexpand2025

Tamper-evident elections with homomorphic encryption preserving voter anonymity.

Approach

Custom Python blockchain with homomorphic encryption for privacy-preserving vote validation and public auditability.

Impact

Verifiable, anonymous elections with a transparent and immutable audit trail.

PythonBlockchainCryptographyHomomorphic Encryption
Crop Recommendation Systemexpand2024

ML system recommending optimal crops using 30 years of agricultural data — 0.84 R².

Approach

Cleaned and merged 30-year soil/climate datasets; trained regression models with Scikit-learn.

Impact

0.84 R² and top-5 crop predictions tailored to soil and climate conditions.

Scikit-learnPandasNumPyRegression Modeling
Smart Home Management Systemexpand2023

Voice-controlled IoT home automation with safety sensors.

Approach

Integrated ESP32 with gas, smoke, and thermal sensors plus irrigation automation.

Impact

Enabled no-code automation for end-users at low cost.

ESP32IoTSensorsHome Automation
CNC Handwriting Machineexpand2023

Two-axis CNC plotter replicating handwriting with sub-millimeter precision.

Approach

G-code interpreter controlling stepper motors on a custom frame with 3D-printed parts.

Impact

Sub-millimeter drawing accuracy across text and freeform paths.

CNCArduinoG-codeStepper Motors3D Printing
Unmanned Person-Following Droneexpand2022

Vision-based quadcopter that autonomously tracks and follows a person.

Approach

Jetson Nano + Pixhawk with real-time visual tracking for stable person-following behavior.

Impact

Stable real-time person-following behavior in open environments.

Jetson NanoPixhawkComputer VisionDrone Control
Collision Avoidance Botexpand2022

Autonomous robot with dynamic obstacle avoidance using proportional control.

Approach

Ultrasonic sensors with proportional control algorithms for real-time speed adjustment.

Impact

Safe adaptive navigation in variable environments.

ArduinoUltrasonic SensorsControl SystemsRobotics
Automatic Result Collection Botexpand2022

Selenium bot for automated academic result scraping and storage.

Approach

Selenium WebDriver with structured storage logic for automated bulk data collection.

Impact

Eliminated manual data collection for large student batches.

PythonSeleniumWeb ScrapingAutomation
Kaggle 2022 MLDS Survey Analysisexpand2022

EDA and Streamlit dashboard on global ML practitioner trends.

Approach

Cleaned and augmented Kaggle survey data; built interactive visualizations in Streamlit.

Impact

Surfaced salary and technology adoption patterns across global ML practitioners.

PandasNumPyMatplotlibSeabornStreamlit

Credentials

Machine Learning Specialization· Stanford Universityexpand2025

Comprehensive specialization covering supervised, unsupervised learning, with recommender systems and deep learning fundamentals.

Three-course specialization taught by Andrew Ng covering regression, classification, neural networks, decision trees, recommender systems, and best practices for ML development. Includes hands-on projects implementing algorithms from scratch.

Mathematics for Machine Learning and Data Science· DeepLearning.AIexpand2024

Comprehensive specialization covering linear algebra, calculus, probability, and statistics.

Focused on building the mathematical foundations required for machine learning, including linear algebra, calculus, and probability & statistics. Included practical exercises applying math concepts to real ML problems.

Education

Master of Engineering in Software Systems

2025 - 2027
Birla Institute of Technology and Science, Pilani
9.34 / 10

Bachelor of Engineering in Electronics and Communication Engineering

2021 - 2025
Osmania University
First Class with Distinction

Voluntary Work

Vice Chair· IEEE Computer Society SB, MJCETexpandJul 2024 - Jun 2025

Assisted the chair in organizing technical events, workshops, and coordinating teams across initiatives.

Impact

Improved event quality and team performance through structured workflows across multiple initiatives.

Vice Chair· The Robotics Centre, MJCETexpandJul 2024 - Jun 2025

Led club operations and strategic planning to grow community engagement and streamline event execution.

Impact

Improved club reach and participation through collaborative initiatives.

Technical Head· The Robotics Centre, MJCETexpandJul 2023 - Jun 2024

Led technical direction for robotics projects, workshops, and competition preparation.

Impact

Executed workshops and achieved strong results at RoboVeda competitions.

Student Chief Editor· Channel ECEexpandDec 2023 - Nov 2024

Led editorial planning and publication of departmental newsletters (annual and monthly).

Impact

Delivered consistent, timely releases under tight deadlines.

Skills

Python Python NumPy NumPy Pandas Pandas Matplotlib Seaborn Scikit-learn Scikit-learn LangChain LangChain LangFlow LangGraph Django Django FastAPI FastAPI PostgreSQL PostgreSQL Docker Docker Kubernetes Kubernetes CI/CD Git Git Bash Bash Agile/Scrum System Design PyTorch PyTorch Jupyter Jupyter Hugging Face Hugging Face MLFlow MLFlow Streamlit Streamlit Flask Flask n8n n8n C C C++ C++