Abhishek Jaju

AI solutions engineer with 3+ years of experience across platform engineering, data systems & applied AI. Built LLM backed evaluation and RAG workflows, production React-Go features and operational dashboards for Uber platform teams. Highly trusted technical owner who sustains a 4.9/5 avg-CSAT score while bridging complex engineering (Python, Go, RAG) with data-driven stakeholder strategy.

Project Engineer at Wipro (Client: Uber) (2024-04 – Present)

• Identified that leadership had no quantitative visibility into api support quality; proposed and executed an LLM evaluation framework unprompted, from problem discovery through production design. Built a multi-stage Python pipeline (preprocessing → metadata extraction → structured scoring → confidence estimation), establishing the first measurable quality baseline and moving coverage from ad-hoc spot-checks to structured scoring, reducing time spent by 40%. Scoped this ambiguous business task directly with stakeholders and iteratively tuned prompts against their feedback to raise evaluation consistency and trust in the output.
• Delivered production React and Go features end-to-end for uCraft, including a user-facing feed card requiring coordinated changes across backend services, IDL contracts and downstream iOS & Android apps.
• Built a RAG based workflow over internal Slack history - retrieving top-k relevant discussions on top of the internal company KBs and injecting them as grounding context - to ground model output in recent discussions rather than parametric recall of documentation. Helped in answering undocumented (tribal knowledge) queries with more than 90% answer accuracy rate across 100 manually evaluated queries.
• Redesigned the dashboard's reporting model with SQL-derived datasets and workflow-aware SLA logic including states such as 'waiting-for-customer', giving leadership visibility into actual resolution timelines. The shipped self-service dashboard (JavaScript, SQL, cloud spreadsheet APIs) replaced an underutilized Tableau, cutting load times by 80%.
• Took over uCraft platform support with minimal documentation with high autonomy; became the primary SME while sustaining a 4.9/5 CSAT across triage, operational decisions, and developer guidance.
• Primary technical contact for 200+ platform requests/month (topic management, consumer troubleshooting, capacity planning) for the Kafka team, part of the Uber Data Platform; maintained zero escalations through a high-demand capacity period. Restructured internal documentation into an LLM-assisted Q&A format, improving discoverability for recurring support scenarios.

Associate Business Analyst at Merkle Sokrati (Dentsu) (2023-05 – 2024-03)

• Translated SQL-driven campaign analysis (ROAS, AOV, CVR) into business recommendations for cross-functional stakeholders, contributing to a 32% increase in Croma's store visitors for targeted retail campaigns.
• Analyzed granular Google Ads performance data daily, identifying underperforming campaigns for key accounts (like Croma, Shoppersstop, Burger King India) and reallocating budget to boost overall efficiency by 25%.

PCM in Pre-University – Heritage International (2018)

Bachelor of Technology in Computer Science Engineering – Deogiri Institute of Engineering & Management Studies (2022)

Nanodegree in Artificial Intelligence Programming with Python – Udacity (2023)