Ahmed Ansari
About Candidate
AI/ML engineer building practical LLM applications, retrieval systems, and developer tools. I work across Python, TypeScript, RAG, MCP, SQL, and agentic workflows, with a focus on reliable systems and measurable outcomes.
Location
Education
Graduated with a 9.15 CGPA. Focused on artificial intelligence, machine learning, data structures, software engineering, databases, and distributed systems. Participated in competitive programming, hackathons, technical meetups, and student technology communities.
Completed a Diploma in Information Technology with a 9.30 CGPA. Built a foundation in programming, web and Android development, software development lifecycle, databases, and core IT systems.
Work & Experience
• Led and built a multi-user AI knowledge assistant over an ITSM product corpus: ~8,500 chunks across 41 modules, hybrid dense and PostgreSQL full-text retrieval, four MCP tool servers, per-user credential isolation, source citations, and ~500 tests. Used daily by eight team members; reduced workflow configuration from 30 minutes–2 hours to ~5 minutes.
• Served as the sole adversarial tester for a shipping LLM feature, covering prompt injection, jailbreaks, scope escape, and client-side exploitation. Around 25 of 40 test cases initially failed; all but five deferred cases were fixed.
• Built a hardware/BOQ sizing tool that reduced sales estimation from ~2 days to ~2 hours, WhatsApp support-ticketing automation serving hundreds of customers daily, a multi-agent content pipeline, and a Text-to-SQL system.
• Designed the programme and content for a company-wide AI enablement initiative reaching ~200 people across seven functions.
Worked in the PMG R&D team on LLM-powered product capabilities for Motadata ServiceOps. Improved natural-language-to-SQL accuracy, response quality, latency, and cost by engineering retrieval and database context for product-facing conversational workflows.
Localized and deployed an open-source LLM for automated question generation. Evaluated Hugging Face models, ran Mistral 7B Instruct in GGUF format through Ollama and LM Studio, and containerized the stack with Docker for Windows, Linux, WSL, and container environments. Also worked with TensorFlow and LangChain.
Completed seven virtual experience programmes spanning API development, AWS solutions architecture, cybersecurity, dashboard automation, data analysis, data science, and software-development infrastructure for organisations including Visa, AWS, Goldman Sachs, Deloitte, British Airways, Quantium, and Accenture.